« back to all changes in this revision
Viewing changes to bzrlib/inventory.py
- browse files at revision 271
- compare with another revision
- download diff
- download tarball
- view history from revision 271
- Committer: Martin Pool
- Date: 2005-04-15 07:53:59 UTC
- Revision ID: mbp@sourcefrog.net-20050415075359-e45b9cdcefc06fc8
- Windows path fixes
- files added:
- bzrlib/remotebranch.py
- files removed:
- 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 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
added
removed
bzrlib/inventory.py
15
15
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
16
16
17
17
18
# TODO: Maybe also keep the full path of the entry, and the children?
19
# But those depend on its position within a particular inventory, and
20
# it would be nice not to need to hold the backpointer here.
18
# TODO: Maybe store inventory_id in the file? Not really needed.
19
21
20
22
21
# This should really be an id randomly assigned when the tree is
23
22
# created, but it's not for now.
24
23
ROOT_ID = "TREE_ROOT"
25
24
26
25
27
import os.path
28
import re
29
import sys
30
import tarfile
31
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
32
36
33
37
import bzrlib
34
from bzrlib.errors import BzrError, BzrCheckError
35
36
from bzrlib.osutils import (pumpfile, quotefn, splitpath, joinpath,
37
appendpath, sha_strings)
38
from bzrlib.osutils import uuid, quotefn, splitpath, joinpath, appendpath
38
39
from bzrlib.trace import mutter
39
from bzrlib.errors import NotVersionedError
40
41
42
class InventoryEntry(object):
40
41
class InventoryEntry(XMLMixin):
43
42
"""Description of a versioned file.
44
43
45
44
An InventoryEntry has the following fields, which are also
46
45
present in the XML inventory-entry element:
47
46
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.)
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
71
61
72
62
>>> i = Inventory()
73
63
>>> i.path2id('')
74
64
'TREE_ROOT'
75
>>> i.add(InventoryDirectory('123', 'src', ROOT_ID))
76
InventoryDirectory('123', 'src', parent_id='TREE_ROOT')
77
>>> i.add(InventoryFile('2323', 'hello.c', parent_id='123'))
78
InventoryFile('2323', 'hello.c', parent_id='123')
65
>>> i.add(InventoryEntry('123', 'src', 'directory', ROOT_ID))
66
>>> i.add(InventoryEntry('2323', 'hello.c', 'file', parent_id='123'))
79
67
>>> for j in i.iter_entries():
80
68
... print j
81
69
...
82
('src', InventoryDirectory('123', 'src', parent_id='TREE_ROOT'))
83
('src/hello.c', InventoryFile('2323', 'hello.c', parent_id='123'))
84
>>> i.add(InventoryFile('2323', 'bye.c', '123'))
70
('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'))
85
73
Traceback (most recent call last):
86
74
...
87
BzrError: inventory already contains entry with id {2323}
88
>>> i.add(InventoryFile('2324', 'bye.c', '123'))
89
InventoryFile('2324', 'bye.c', parent_id='123')
90
>>> i.add(InventoryDirectory('2325', 'wibble', '123'))
91
InventoryDirectory('2325', 'wibble', parent_id='123')
75
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'))
92
78
>>> i.path2id('src/wibble')
93
79
'2325'
94
80
>>> '2325' in i
95
81
True
96
>>> i.add(InventoryFile('2326', 'wibble.c', '2325'))
97
InventoryFile('2326', 'wibble.c', parent_id='2325')
82
>>> i.add(InventoryEntry('2326', 'wibble.c', 'file', '2325'))
98
83
>>> i['2326']
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)
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])
103
88
...
104
89
src
105
90
src/bye.c
106
91
src/hello.c
107
92
src/wibble
108
93
src/wibble/wibble.c
109
>>> i.id2path('2326').replace('\\\\', '/')
94
>>> i.id2path('2326')
110
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.
111
100
"""
101
102
# TODO: split InventoryEntry into subclasses for files,
103
# directories, etc etc.
112
104
113
__slots__ = ['text_sha1', 'text_size', 'file_id', 'name', 'kind',
114
'text_id', 'parent_id', 'children', 'executable',
115
'revision']
116
117
def _add_text_to_weave(self, new_lines, parents, weave_store):
118
weave_store.add_text(self.file_id, self.revision, new_lines, parents)
119
120
def detect_changes(self, old_entry):
121
"""Return a (text_modified, meta_modified) from this to old_entry.
122
123
_read_tree_state must have been called on self and old_entry prior to
124
calling detect_changes.
125
"""
126
return False, False
127
128
def diff(self, text_diff, from_label, tree, to_label, to_entry, to_tree,
129
output_to, reverse=False):
130
"""Perform a diff from this to to_entry.
131
132
text_diff will be used for textual difference calculation.
133
This is a template method, override _diff in child classes.
134
"""
135
self._read_tree_state(tree.id2path(self.file_id), tree)
136
if to_entry:
137
# cannot diff from one kind to another - you must do a removal
138
# and an addif they do not match.
139
assert self.kind == to_entry.kind
140
to_entry._read_tree_state(to_tree.id2path(to_entry.file_id),
141
to_tree)
142
self._diff(text_diff, from_label, tree, to_label, to_entry, to_tree,
143
output_to, reverse)
144
145
def _diff(self, text_diff, from_label, tree, to_label, to_entry, to_tree,
146
output_to, reverse=False):
147
"""Perform a diff between two entries of the same kind."""
148
149
def find_previous_heads(self, previous_inventories, entry_weave):
150
"""Return the revisions and entries that directly preceed this.
151
152
Returned as a map from revision to inventory entry.
153
154
This is a map containing the file revisions in all parents
155
for which the file exists, and its revision is not a parent of
156
any other. If the file is new, the set will be empty.
157
"""
158
def get_ancestors(weave, entry):
159
return set(map(weave.idx_to_name,
160
weave.inclusions([weave.lookup(entry.revision)])))
161
heads = {}
162
head_ancestors = {}
163
for inv in previous_inventories:
164
if self.file_id in inv:
165
ie = inv[self.file_id]
166
assert ie.file_id == self.file_id
167
if ie.revision in heads:
168
assert heads[ie.revision] == ie
169
else:
170
# may want to add it.
171
# may already be covered:
172
already_present = 0 != len(
173
[head for head in heads
174
if ie.revision in head_ancestors[head]])
175
if already_present:
176
# an ancestor of a known head.
177
continue
178
# definately a head:
179
ancestors = get_ancestors(entry_weave, ie)
180
# may knock something else out:
181
check_heads = list(heads.keys())
182
for head in check_heads:
183
if head in ancestors:
184
# this head is not really a head
185
heads.pop(head)
186
head_ancestors[ie.revision] = ancestors
187
heads[ie.revision] = ie
188
return heads
189
190
def get_tar_item(self, root, dp, now, tree):
191
"""Get a tarfile item and a file stream for its content."""
192
item = tarfile.TarInfo(os.path.join(root, dp))
193
# TODO: would be cool to actually set it to the timestamp of the
194
# revision it was last changed
195
item.mtime = now
196
fileobj = self._put_in_tar(item, tree)
197
return item, fileobj
198
199
def has_text(self):
200
"""Return true if the object this entry represents has textual data.
201
202
Note that textual data includes binary content.
203
204
Also note that all entries get weave files created for them.
205
This attribute is primarily used when upgrading from old trees that
206
did not have the weave index for all inventory entries.
207
"""
208
return False
209
210
def __init__(self, file_id, name, parent_id, text_id=None):
105
def __init__(self, file_id, name, kind, parent_id, text_id=None):
211
106
"""Create an InventoryEntry
212
107
213
108
The filename must be a single component, relative to the
214
109
parent directory; it cannot be a whole path or relative name.
215
110
216
>>> e = InventoryFile('123', 'hello.c', ROOT_ID)
111
>>> e = InventoryEntry('123', 'hello.c', 'file', ROOT_ID)
217
112
>>> e.name
218
113
'hello.c'
219
114
>>> e.file_id
220
115
'123'
221
>>> e = InventoryFile('123', 'src/hello.c', ROOT_ID)
116
>>> e = InventoryEntry('123', 'src/hello.c', 'file', ROOT_ID)
222
117
Traceback (most recent call last):
223
BzrCheckError: InventoryEntry name 'src/hello.c' is invalid
118
BzrError: ("InventoryEntry name is not a simple filename: 'src/hello.c'", [])
224
119
"""
225
assert isinstance(name, basestring), name
226
if '/' in name or '\\' in name:
227
raise BzrCheckError('InventoryEntry name %r is invalid' % name)
228
229
self.executable = False
230
self.revision = None
231
self.text_sha1 = None
232
self.text_size = None
120
121
if len(splitpath(name)) != 1:
122
bailout('InventoryEntry name is not a simple filename: %r'
123
% name)
124
233
125
self.file_id = file_id
234
126
self.name = name
127
self.kind = kind
235
128
self.text_id = text_id
236
129
self.parent_id = parent_id
237
self.symlink_target = None
238
239
def kind_character(self):
240
"""Return a short kind indicator useful for appending to names."""
241
raise BzrError('unknown kind %r' % self.kind)
242
243
known_kinds = ('file', 'directory', 'symlink', 'root_directory')
244
245
def _put_in_tar(self, item, tree):
246
"""populate item for stashing in a tar, and return the content stream.
247
248
If no content is available, return None.
249
"""
250
raise BzrError("don't know how to export {%s} of kind %r" %
251
(self.file_id, self.kind))
252
253
def put_on_disk(self, dest, dp, tree):
254
"""Create a representation of self on disk in the prefix dest.
255
256
This is a template method - implement _put_on_disk in subclasses.
257
"""
258
fullpath = appendpath(dest, dp)
259
self._put_on_disk(fullpath, tree)
260
mutter(" export {%s} kind %s to %s" % (self.file_id, self.kind, fullpath))
261
262
def _put_on_disk(self, fullpath, tree):
263
"""Put this entry onto disk at fullpath, from tree tree."""
264
raise BzrError("don't know how to export {%s} of kind %r" % (self.file_id, self.kind))
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
265
140
266
141
def sorted_children(self):
267
142
l = self.children.items()
268
143
l.sort()
269
144
return l
270
145
271
@staticmethod
272
def versionable_kind(kind):
273
return kind in ('file', 'directory', 'symlink')
274
275
def check(self, checker, rev_id, inv, tree):
276
"""Check this inventory entry is intact.
277
278
This is a template method, override _check for kind specific
279
tests.
280
"""
281
if self.parent_id != None:
282
if not inv.has_id(self.parent_id):
283
raise BzrCheckError('missing parent {%s} in inventory for revision {%s}'
284
% (self.parent_id, rev_id))
285
self._check(checker, rev_id, tree)
286
287
def _check(self, checker, rev_id, tree):
288
"""Check this inventory entry for kind specific errors."""
289
raise BzrCheckError('unknown entry kind %r in revision {%s}' %
290
(self.kind, rev_id))
291
292
146
293
147
def copy(self):
294
"""Clone this inventory entry."""
295
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
296
153
297
def _get_snapshot_change(self, previous_entries):
298
if len(previous_entries) > 1:
299
return 'merged'
300
elif len(previous_entries) == 0:
301
return 'added'
302
else:
303
return 'modified/renamed/reparented'
304
154
305
155
def __repr__(self):
306
return ("%s(%r, %r, parent_id=%r)"
156
return ("%s(%r, %r, kind=%r, parent_id=%r)"
307
157
% (self.__class__.__name__,
308
158
self.file_id,
309
159
self.name,
160
self.kind,
310
161
self.parent_id))
311
162
312
def snapshot(self, revision, path, previous_entries,
313
work_tree, weave_store):
314
"""Make a snapshot of this entry which may or may not have changed.
315
316
This means that all its fields are populated, that it has its
317
text stored in the text store or weave.
318
"""
319
mutter('new parents of %s are %r', path, previous_entries)
320
self._read_tree_state(path, work_tree)
321
if len(previous_entries) == 1:
322
# cannot be unchanged unless there is only one parent file rev.
323
parent_ie = previous_entries.values()[0]
324
if self._unchanged(parent_ie):
325
mutter("found unchanged entry")
326
self.revision = parent_ie.revision
327
return "unchanged"
328
return self.snapshot_revision(revision, previous_entries,
329
work_tree, weave_store)
330
331
def snapshot_revision(self, revision, previous_entries, work_tree,
332
weave_store):
333
"""Record this revision unconditionally."""
334
mutter('new revision for {%s}', self.file_id)
335
self.revision = revision
336
change = self._get_snapshot_change(previous_entries)
337
self._snapshot_text(previous_entries, work_tree, weave_store)
338
return change
339
340
def _snapshot_text(self, file_parents, work_tree, weave_store):
341
"""Record the 'text' of this entry, whatever form that takes.
342
343
This default implementation simply adds an empty text.
344
"""
345
mutter('storing file {%s} in revision {%s}',
346
self.file_id, self.revision)
347
self._add_text_to_weave([], file_parents, weave_store)
348
349
def __eq__(self, other):
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
350
219
if not isinstance(other, InventoryEntry):
351
220
return NotImplemented
352
221
353
return ((self.file_id == other.file_id)
354
and (self.name == other.name)
355
and (other.symlink_target == self.symlink_target)
356
and (self.text_sha1 == other.text_sha1)
357
and (self.text_size == other.text_size)
358
and (self.text_id == other.text_id)
359
and (self.parent_id == other.parent_id)
360
and (self.kind == other.kind)
361
and (self.revision == other.revision)
362
and (self.executable == other.executable)
363
)
364
365
def __ne__(self, other):
366
return not (self == other)
367
368
def __hash__(self):
369
raise ValueError('not hashable')
370
371
def _unchanged(self, previous_ie):
372
"""Has this entry changed relative to previous_ie.
373
374
This method should be overriden in child classes.
375
"""
376
compatible = True
377
# different inv parent
378
if previous_ie.parent_id != self.parent_id:
379
compatible = False
380
# renamed
381
elif previous_ie.name != self.name:
382
compatible = False
383
return compatible
384
385
def _read_tree_state(self, path, work_tree):
386
"""Populate fields in the inventory entry from the given tree.
387
388
Note that this should be modified to be a noop on virtual trees
389
as all entries created there are prepopulated.
390
"""
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
391
230
392
231
393
232
class RootEntry(InventoryEntry):
394
395
def _check(self, checker, rev_id, tree):
396
"""See InventoryEntry._check"""
397
398
233
def __init__(self, file_id):
399
234
self.file_id = file_id
400
235
self.children = {}
402
237
self.parent_id = None
403
238
self.name = ''
404
239
405
def __eq__(self, other):
240
def __cmp__(self, other):
241
if self is other:
242
return 0
406
243
if not isinstance(other, RootEntry):
407
244
return NotImplemented
408
409
return (self.file_id == other.file_id) \
410
and (self.children == other.children)
411
412
413
class InventoryDirectory(InventoryEntry):
414
"""A directory in an inventory."""
415
416
def _check(self, checker, rev_id, tree):
417
"""See InventoryEntry._check"""
418
if self.text_sha1 != None or self.text_size != None or self.text_id != None:
419
raise BzrCheckError('directory {%s} has text in revision {%s}'
420
% (self.file_id, rev_id))
421
422
def copy(self):
423
other = InventoryDirectory(self.file_id, self.name, self.parent_id)
424
other.revision = self.revision
425
# note that children are *not* copied; they're pulled across when
426
# others are added
427
return other
428
429
def __init__(self, file_id, name, parent_id):
430
super(InventoryDirectory, self).__init__(file_id, name, parent_id)
431
self.children = {}
432
self.kind = 'directory'
433
434
def kind_character(self):
435
"""See InventoryEntry.kind_character."""
436
return '/'
437
438
def _put_in_tar(self, item, tree):
439
"""See InventoryEntry._put_in_tar."""
440
item.type = tarfile.DIRTYPE
441
fileobj = None
442
item.name += '/'
443
item.size = 0
444
item.mode = 0755
445
return fileobj
446
447
def _put_on_disk(self, fullpath, tree):
448
"""See InventoryEntry._put_on_disk."""
449
os.mkdir(fullpath)
450
451
452
class InventoryFile(InventoryEntry):
453
"""A file in an inventory."""
454
455
def _check(self, checker, rev_id, tree):
456
"""See InventoryEntry._check"""
457
revision = self.revision
458
t = (self.file_id, revision)
459
if t in checker.checked_texts:
460
prev_sha = checker.checked_texts[t]
461
if prev_sha != self.text_sha1:
462
raise BzrCheckError('mismatched sha1 on {%s} in {%s}' %
463
(self.file_id, rev_id))
464
else:
465
checker.repeated_text_cnt += 1
466
return
467
mutter('check version {%s} of {%s}', rev_id, self.file_id)
468
file_lines = tree.get_file_lines(self.file_id)
469
checker.checked_text_cnt += 1
470
if self.text_size != sum(map(len, file_lines)):
471
raise BzrCheckError('text {%s} wrong size' % self.text_id)
472
if self.text_sha1 != sha_strings(file_lines):
473
raise BzrCheckError('text {%s} wrong sha1' % self.text_id)
474
checker.checked_texts[t] = self.text_sha1
475
476
def copy(self):
477
other = InventoryFile(self.file_id, self.name, self.parent_id)
478
other.executable = self.executable
479
other.text_id = self.text_id
480
other.text_sha1 = self.text_sha1
481
other.text_size = self.text_size
482
other.revision = self.revision
483
return other
484
485
def detect_changes(self, old_entry):
486
"""See InventoryEntry.detect_changes."""
487
assert self.text_sha1 != None
488
assert old_entry.text_sha1 != None
489
text_modified = (self.text_sha1 != old_entry.text_sha1)
490
meta_modified = (self.executable != old_entry.executable)
491
return text_modified, meta_modified
492
493
def _diff(self, text_diff, from_label, tree, to_label, to_entry, to_tree,
494
output_to, reverse=False):
495
"""See InventoryEntry._diff."""
496
from_text = tree.get_file(self.file_id).readlines()
497
if to_entry:
498
to_text = to_tree.get_file(to_entry.file_id).readlines()
499
else:
500
to_text = []
501
if not reverse:
502
text_diff(from_label, from_text,
503
to_label, to_text, output_to)
504
else:
505
text_diff(to_label, to_text,
506
from_label, from_text, output_to)
507
508
def has_text(self):
509
"""See InventoryEntry.has_text."""
510
return True
511
512
def __init__(self, file_id, name, parent_id):
513
super(InventoryFile, self).__init__(file_id, name, parent_id)
514
self.kind = 'file'
515
516
def kind_character(self):
517
"""See InventoryEntry.kind_character."""
518
return ''
519
520
def _put_in_tar(self, item, tree):
521
"""See InventoryEntry._put_in_tar."""
522
item.type = tarfile.REGTYPE
523
fileobj = tree.get_file(self.file_id)
524
item.size = self.text_size
525
if tree.is_executable(self.file_id):
526
item.mode = 0755
527
else:
528
item.mode = 0644
529
return fileobj
530
531
def _put_on_disk(self, fullpath, tree):
532
"""See InventoryEntry._put_on_disk."""
533
pumpfile(tree.get_file(self.file_id), file(fullpath, 'wb'))
534
if tree.is_executable(self.file_id):
535
os.chmod(fullpath, 0755)
536
537
def _read_tree_state(self, path, work_tree):
538
"""See InventoryEntry._read_tree_state."""
539
self.text_sha1 = work_tree.get_file_sha1(self.file_id)
540
self.executable = work_tree.is_executable(self.file_id)
541
542
def _snapshot_text(self, file_parents, work_tree, weave_store):
543
"""See InventoryEntry._snapshot_text."""
544
mutter('storing file {%s} in revision {%s}',
545
self.file_id, self.revision)
546
# special case to avoid diffing on renames or
547
# reparenting
548
if (len(file_parents) == 1
549
and self.text_sha1 == file_parents.values()[0].text_sha1
550
and self.text_size == file_parents.values()[0].text_size):
551
previous_ie = file_parents.values()[0]
552
weave_store.add_identical_text(
553
self.file_id, previous_ie.revision,
554
self.revision, file_parents)
555
else:
556
new_lines = work_tree.get_file(self.file_id).readlines()
557
self._add_text_to_weave(new_lines, file_parents, weave_store)
558
self.text_sha1 = sha_strings(new_lines)
559
self.text_size = sum(map(len, new_lines))
560
561
562
def _unchanged(self, previous_ie):
563
"""See InventoryEntry._unchanged."""
564
compatible = super(InventoryFile, self)._unchanged(previous_ie)
565
if self.text_sha1 != previous_ie.text_sha1:
566
compatible = False
567
else:
568
# FIXME: 20050930 probe for the text size when getting sha1
569
# in _read_tree_state
570
self.text_size = previous_ie.text_size
571
return compatible
572
573
574
class InventoryLink(InventoryEntry):
575
"""A file in an inventory."""
576
577
__slots__ = ['symlink_target']
578
579
def _check(self, checker, rev_id, tree):
580
"""See InventoryEntry._check"""
581
if self.text_sha1 != None or self.text_size != None or self.text_id != None:
582
raise BzrCheckError('symlink {%s} has text in revision {%s}'
583
% (self.file_id, rev_id))
584
if self.symlink_target == None:
585
raise BzrCheckError('symlink {%s} has no target in revision {%s}'
586
% (self.file_id, rev_id))
587
588
def copy(self):
589
other = InventoryLink(self.file_id, self.name, self.parent_id)
590
other.symlink_target = self.symlink_target
591
other.revision = self.revision
592
return other
593
594
def detect_changes(self, old_entry):
595
"""See InventoryEntry.detect_changes."""
596
# FIXME: which _modified field should we use ? RBC 20051003
597
text_modified = (self.symlink_target != old_entry.symlink_target)
598
if text_modified:
599
mutter(" symlink target changed")
600
meta_modified = False
601
return text_modified, meta_modified
602
603
def _diff(self, text_diff, from_label, tree, to_label, to_entry, to_tree,
604
output_to, reverse=False):
605
"""See InventoryEntry._diff."""
606
from_text = self.symlink_target
607
if to_entry is not None:
608
to_text = to_entry.symlink_target
609
if reverse:
610
temp = from_text
611
from_text = to_text
612
to_text = temp
613
print >>output_to, '=== target changed %r => %r' % (from_text, to_text)
614
else:
615
if not reverse:
616
print >>output_to, '=== target was %r' % self.symlink_target
617
else:
618
print >>output_to, '=== target is %r' % self.symlink_target
619
620
def __init__(self, file_id, name, parent_id):
621
super(InventoryLink, self).__init__(file_id, name, parent_id)
622
self.kind = 'symlink'
623
624
def kind_character(self):
625
"""See InventoryEntry.kind_character."""
626
return ''
627
628
def _put_in_tar(self, item, tree):
629
"""See InventoryEntry._put_in_tar."""
630
iterm.type = tarfile.SYMTYPE
631
fileobj = None
632
item.size = 0
633
item.mode = 0755
634
item.linkname = self.symlink_target
635
return fileobj
636
637
def _put_on_disk(self, fullpath, tree):
638
"""See InventoryEntry._put_on_disk."""
639
try:
640
os.symlink(self.symlink_target, fullpath)
641
except OSError,e:
642
raise BzrError("Failed to create symlink %r -> %r, error: %s" % (fullpath, self.symlink_target, e))
643
644
def _read_tree_state(self, path, work_tree):
645
"""See InventoryEntry._read_tree_state."""
646
self.symlink_target = work_tree.get_symlink_target(self.file_id)
647
648
def _unchanged(self, previous_ie):
649
"""See InventoryEntry._unchanged."""
650
compatible = super(InventoryLink, self)._unchanged(previous_ie)
651
if self.symlink_target != previous_ie.symlink_target:
652
compatible = False
653
return compatible
654
655
656
class Inventory(object):
245
return cmp(self.file_id, other.file_id) \
246
or cmp(self.children, other.children)
247
248
249
250
class Inventory(XMLMixin):
657
251
"""Inventory of versioned files in a tree.
658
252
659
253
This describes which file_id is present at each point in the tree,
671
265
inserted, other than through the Inventory API.
672
266
673
267
>>> inv = Inventory()
674
>>> inv.add(InventoryFile('123-123', 'hello.c', ROOT_ID))
675
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))
676
272
>>> inv['123-123'].name
677
273
'hello.c'
678
274
687
283
688
284
>>> [x[0] for x in inv.iter_entries()]
689
285
['hello.c']
690
>>> inv = Inventory('TREE_ROOT-12345678-12345678')
691
>>> inv.add(InventoryFile('123-123', 'hello.c', ROOT_ID))
692
InventoryFile('123-123', 'hello.c', parent_id='TREE_ROOT-12345678-12345678')
286
287
>>> inv.write_xml(sys.stdout)
288
<inventory>
289
<entry file_id="123-123" kind="file" name="hello.c" />
290
</inventory>
291
693
292
"""
694
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):
695
310
"""Create or read an inventory.
696
311
697
312
If a working directory is specified, the inventory is read
701
316
The inventory is created with a default root directory, with
702
317
an id of None.
703
318
"""
704
# We are letting Branch.initialize() create a unique inventory
705
# root id. Rather than generating a random one here.
706
#if root_id is None:
707
# root_id = bzrlib.branch.gen_file_id('TREE_ROOT')
708
self.root = RootEntry(root_id)
319
self.root = RootEntry(ROOT_ID)
709
320
self._byid = {self.root.file_id: self.root}
710
321
711
322
712
def copy(self):
713
other = Inventory(self.root.file_id)
714
# copy recursively so we know directories will be added before
715
# their children. There are more efficient ways than this...
716
for path, entry in self.iter_entries():
717
if entry == self.root:
718
continue
719
other.add(entry.copy())
720
return other
721
722
723
323
def __iter__(self):
724
324
return iter(self._byid)
725
325
744
344
if ie.kind == 'directory':
745
345
for cn, cie in self.iter_entries(from_dir=ie.file_id):
746
346
yield os.path.join(name, cn), cie
747
748
749
def entries(self):
750
"""Return list of (path, ie) for all entries except the root.
751
752
This may be faster than iter_entries.
347
348
349
350
def directories(self):
351
"""Return (path, entry) pairs for all directories.
753
352
"""
754
accum = []
755
def descend(dir_ie, dir_path):
756
kids = dir_ie.children.items()
757
kids.sort()
758
for name, ie in kids:
759
child_path = os.path.join(dir_path, name)
760
accum.append((child_path, ie))
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():
761
360
if ie.kind == 'directory':
762
descend(ie, child_path)
763
764
descend(self.root, '')
765
return accum
766
767
768
def directories(self):
769
"""Return (path, entry) pairs for all directories, including the root.
770
"""
771
accum = []
772
def descend(parent_ie, parent_path):
773
accum.append((parent_path, parent_ie))
361
dn.append((ie.name, ie))
362
dn.sort()
774
363
775
kids = [(ie.name, ie) for ie in parent_ie.children.itervalues() if ie.kind == 'directory']
776
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
777
367
778
for name, child_ie in kids:
779
child_path = os.path.join(parent_path, name)
780
descend(child_ie, child_path)
781
descend(self.root, '')
782
return accum
368
for name, ie in descend(self.root):
369
yield name, ie
783
370
784
371
785
372
787
374
"""True if this entry contains a file with given id.
788
375
789
376
>>> inv = Inventory()
790
>>> inv.add(InventoryFile('123', 'foo.c', ROOT_ID))
791
InventoryFile('123', 'foo.c', parent_id='TREE_ROOT')
377
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
792
378
>>> '123' in inv
793
379
True
794
380
>>> '456' in inv
801
387
"""Return the entry for given file_id.
802
388
803
389
>>> inv = Inventory()
804
>>> inv.add(InventoryFile('123123', 'hello.c', ROOT_ID))
805
InventoryFile('123123', 'hello.c', parent_id='TREE_ROOT')
390
>>> inv.add(InventoryEntry('123123', 'hello.c', 'file', ROOT_ID))
806
391
>>> inv['123123'].name
807
392
'hello.c'
808
393
"""
394
if file_id == None:
395
raise BzrError("can't look up file_id None")
396
809
397
try:
810
398
return self._byid[file_id]
811
399
except KeyError:
812
if file_id == None:
813
raise BzrError("can't look up file_id None")
814
else:
815
raise BzrError("file_id {%s} not in inventory" % file_id)
816
817
818
def get_file_kind(self, file_id):
819
return self._byid[file_id].kind
400
raise BzrError("file_id {%s} not in inventory" % file_id)
401
820
402
821
403
def get_child(self, parent_id, filename):
822
404
return self[parent_id].children.get(filename)
826
408
"""Add entry to inventory.
827
409
828
410
To add a file to a branch ready to be committed, use Branch.add,
829
which calls this.
830
831
Returns the new entry object.
832
"""
411
which calls this."""
833
412
if entry.file_id in self._byid:
834
raise BzrError("inventory already contains entry with id {%s}" % entry.file_id)
835
836
if entry.parent_id == ROOT_ID or entry.parent_id is None:
837
entry.parent_id = self.root.file_id
413
bailout("inventory already contains entry with id {%s}" % entry.file_id)
838
414
839
415
try:
840
416
parent = self._byid[entry.parent_id]
841
417
except KeyError:
842
raise BzrError("parent_id {%s} not in inventory" % entry.parent_id)
418
bailout("parent_id {%s} not in inventory" % entry.parent_id)
843
419
844
420
if parent.children.has_key(entry.name):
845
raise BzrError("%s is already versioned" %
421
bailout("%s is already versioned" %
846
422
appendpath(self.id2path(parent.file_id), entry.name))
847
423
848
424
self._byid[entry.file_id] = entry
849
425
parent.children[entry.name] = entry
850
return entry
851
426
852
427
853
428
def add_path(self, relpath, kind, file_id=None):
854
429
"""Add entry from a path.
855
430
856
The immediate parent must already be versioned.
857
858
Returns the new entry object."""
859
from bzrlib.branch import gen_file_id
860
431
The immediate parent must already be versioned"""
861
432
parts = bzrlib.osutils.splitpath(relpath)
862
433
if len(parts) == 0:
863
raise BzrError("cannot re-add root of inventory")
434
bailout("cannot re-add root of inventory")
864
435
865
436
if file_id == None:
866
file_id = gen_file_id(relpath)
867
868
parent_path = parts[:-1]
869
parent_id = self.path2id(parent_path)
870
if parent_id == None:
871
raise NotVersionedError(parent_path)
872
873
if kind == 'directory':
874
ie = InventoryDirectory(file_id, parts[-1], parent_id)
875
elif kind == 'file':
876
ie = InventoryFile(file_id, parts[-1], parent_id)
877
elif kind == 'symlink':
878
ie = InventoryLink(file_id, parts[-1], parent_id)
879
else:
880
raise BzrError("unknown kind %r" % kind)
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)
881
443
return self.add(ie)
882
444
883
445
885
447
"""Remove entry by id.
886
448
887
449
>>> inv = Inventory()
888
>>> inv.add(InventoryFile('123', 'foo.c', ROOT_ID))
889
InventoryFile('123', 'foo.c', parent_id='TREE_ROOT')
450
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
890
451
>>> '123' in inv
891
452
True
892
453
>>> del inv['123']
908
469
del self[ie.parent_id].children[ie.name]
909
470
910
471
911
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):
912
505
"""Compare two sets by comparing their contents.
913
506
914
507
>>> i1 = Inventory()
915
508
>>> i2 = Inventory()
916
509
>>> i1 == i2
917
510
True
918
>>> i1.add(InventoryFile('123', 'foo', ROOT_ID))
919
InventoryFile('123', 'foo', parent_id='TREE_ROOT')
511
>>> i1.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
920
512
>>> i1 == i2
921
513
False
922
>>> i2.add(InventoryFile('123', 'foo', ROOT_ID))
923
InventoryFile('123', 'foo', parent_id='TREE_ROOT')
514
>>> i2.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
924
515
>>> i1 == i2
925
516
True
926
517
"""
518
if self is other:
519
return 0
520
927
521
if not isinstance(other, Inventory):
928
522
return NotImplemented
929
523
930
if len(self._byid) != len(other._byid):
931
# shortcut: obviously not the same
932
return False
933
934
return self._byid == other._byid
935
936
937
def __ne__(self, other):
938
return not self.__eq__(other)
939
940
941
def __hash__(self):
942
raise ValueError('not hashable')
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
943
532
944
533
945
534
def get_idpath(self, file_id):
955
544
try:
956
545
ie = self._byid[file_id]
957
546
except KeyError:
958
raise BzrError("file_id {%s} not found in inventory" % file_id)
547
bailout("file_id {%s} not found in inventory" % file_id)
959
548
p.insert(0, ie.file_id)
960
549
file_id = ie.parent_id
961
550
return p
962
551
963
552
964
553
def id2path(self, file_id):
965
"""Return as a list the path to file_id.
966
967
>>> i = Inventory()
968
>>> e = i.add(InventoryDirectory('src-id', 'src', ROOT_ID))
969
>>> e = i.add(InventoryFile('foo-id', 'foo.c', parent_id='src-id'))
970
>>> print i.id2path('foo-id').replace(os.sep, '/')
971
src/foo.c
972
"""
554
"""Return as a list the path to file_id."""
555
973
556
# get all names, skipping root
974
p = [self._byid[fid].name for fid in self.get_idpath(file_id)[1:]]
557
p = [self[fid].name for fid in self.get_idpath(file_id)[1:]]
975
558
return os.sep.join(p)
976
559
977
560
1021
604
1022
605
This does not move the working file."""
1023
606
if not is_valid_name(new_name):
1024
raise BzrError("not an acceptable filename: %r" % new_name)
607
bailout("not an acceptable filename: %r" % new_name)
1025
608
1026
609
new_parent = self._byid[new_parent_id]
1027
610
if new_name in new_parent.children:
1028
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)))
1029
612
1030
613
new_parent_idpath = self.get_idpath(new_parent_id)
1031
614
if file_id in new_parent_idpath:
1032
raise BzrError("cannot move directory %r into a subdirectory of itself, %r"
615
bailout("cannot move directory %r into a subdirectory of itself, %r"
1033
616
% (self.id2path(file_id), self.id2path(new_parent_id)))
1034
617
1035
618
file_ie = self._byid[file_id]
1046
629
1047
630
1048
631
1049
_NAME_RE = None
632
_NAME_RE = re.compile(r'^[^/\\]+$')
1050
633
1051
634
def is_valid_name(name):
1052
global _NAME_RE
1053
if _NAME_RE == None:
1054
_NAME_RE = re.compile(r'^[^/\\]+$')
1055
1056
635
return bool(_NAME_RE.match(name))
Loggerhead is a web-based interface for Breezy
Version: 2.0.1