« back to all changes in this revision
Viewing changes to bzrlib/inventory.py
- browse files at revision 316
- compare with another revision
- download diff
- download tarball
- view history from revision 316
- Committer: Martin Pool
- Date: 2005-05-02 04:24:33 UTC
- Revision ID: mbp@sourcefrog.net-20050502042433-c825a7f7235f6b15
doc: notes on merge
- 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
- bzr *
- setup.py *
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, 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):
105
def __init__(self, file_id, name, kind, parent_id, text_id=None):
220
106
"""Create an InventoryEntry
221
107
222
108
The filename must be a single component, relative to the
223
109
parent directory; it cannot be a whole path or relative name.
224
110
225
>>> e = InventoryFile('123', 'hello.c', ROOT_ID)
111
>>> e = InventoryEntry('123', 'hello.c', 'file', ROOT_ID)
226
112
>>> e.name
227
113
'hello.c'
228
114
>>> e.file_id
229
115
'123'
230
>>> e = InventoryFile('123', 'src/hello.c', ROOT_ID)
116
>>> e = InventoryEntry('123', 'src/hello.c', 'file', ROOT_ID)
231
117
Traceback (most recent call last):
232
BzrCheckError: InventoryEntry name 'src/hello.c' is invalid
118
BzrError: ("InventoryEntry name is not a simple filename: 'src/hello.c'", [])
233
119
"""
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
120
121
if len(splitpath(name)) != 1:
122
bailout('InventoryEntry name is not a simple filename: %r'
123
% name)
124
242
125
self.file_id = file_id
243
126
self.name = name
127
self.kind = kind
244
128
self.text_id = text_id
245
129
self.parent_id = parent_id
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))
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
274
140
275
141
def sorted_children(self):
276
142
l = self.children.items()
277
143
l.sort()
278
144
return l
279
145
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
301
146
302
147
def copy(self):
303
"""Clone this inventory entry."""
304
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
305
153
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'
313
154
314
155
def __repr__(self):
315
return ("%s(%r, %r, parent_id=%r)"
156
return ("%s(%r, %r, kind=%r, parent_id=%r)"
316
157
% (self.__class__.__name__,
317
158
self.file_id,
318
159
self.name,
160
self.kind,
319
161
self.parent_id))
320
162
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):
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
360
219
if not isinstance(other, InventoryEntry):
361
220
return NotImplemented
362
221
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
"""
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
401
230
402
231
403
232
class RootEntry(InventoryEntry):
404
405
def _check(self, checker, rev_id, tree):
406
"""See InventoryEntry._check"""
407
408
233
def __init__(self, file_id):
409
234
self.file_id = file_id
410
235
self.children = {}
412
237
self.parent_id = None
413
238
self.name = ''
414
239
415
def __eq__(self, other):
240
def __cmp__(self, other):
241
if self is other:
242
return 0
416
243
if not isinstance(other, RootEntry):
417
244
return NotImplemented
418
419
return (self.file_id == other.file_id) \
420
and (self.children == other.children)
421
422
423
class InventoryDirectory(InventoryEntry):
424
"""A directory in an inventory."""
425
426
def _check(self, checker, rev_id, tree):
427
"""See InventoryEntry._check"""
428
if self.text_sha1 != None or self.text_size != None or self.text_id != None:
429
raise BzrCheckError('directory {%s} has text in revision {%s}'
430
% (self.file_id, rev_id))
431
432
def copy(self):
433
other = InventoryDirectory(self.file_id, self.name, self.parent_id)
434
other.revision = self.revision
435
# note that children are *not* copied; they're pulled across when
436
# others are added
437
return other
438
439
def __init__(self, file_id, name, parent_id):
440
super(InventoryDirectory, self).__init__(file_id, name, parent_id)
441
self.children = {}
442
self.kind = 'directory'
443
444
def kind_character(self):
445
"""See InventoryEntry.kind_character."""
446
return '/'
447
448
def _put_in_tar(self, item, tree):
449
"""See InventoryEntry._put_in_tar."""
450
item.type = tarfile.DIRTYPE
451
fileobj = None
452
item.name += '/'
453
item.size = 0
454
item.mode = 0755
455
return fileobj
456
457
def _put_on_disk(self, fullpath, tree):
458
"""See InventoryEntry._put_on_disk."""
459
os.mkdir(fullpath)
460
461
462
class InventoryFile(InventoryEntry):
463
"""A file in an inventory."""
464
465
def _check(self, checker, rev_id, tree):
466
"""See InventoryEntry._check"""
467
revision = self.revision
468
t = (self.file_id, revision)
469
if t in checker.checked_texts:
470
prev_sha = checker.checked_texts[t]
471
if prev_sha != self.text_sha1:
472
raise BzrCheckError('mismatched sha1 on {%s} in {%s}' %
473
(self.file_id, rev_id))
474
else:
475
checker.repeated_text_cnt += 1
476
return
477
mutter('check version {%s} of {%s}', rev_id, self.file_id)
478
file_lines = tree.get_file_lines(self.file_id)
479
checker.checked_text_cnt += 1
480
if self.text_size != sum(map(len, file_lines)):
481
raise BzrCheckError('text {%s} wrong size' % self.text_id)
482
if self.text_sha1 != sha_strings(file_lines):
483
raise BzrCheckError('text {%s} wrong sha1' % self.text_id)
484
checker.checked_texts[t] = self.text_sha1
485
486
def copy(self):
487
other = InventoryFile(self.file_id, self.name, self.parent_id)
488
other.executable = self.executable
489
other.text_id = self.text_id
490
other.text_sha1 = self.text_sha1
491
other.text_size = self.text_size
492
other.revision = self.revision
493
return other
494
495
def detect_changes(self, old_entry):
496
"""See InventoryEntry.detect_changes."""
497
assert self.text_sha1 != None
498
assert old_entry.text_sha1 != None
499
text_modified = (self.text_sha1 != old_entry.text_sha1)
500
meta_modified = (self.executable != old_entry.executable)
501
return text_modified, meta_modified
502
503
def _diff(self, text_diff, from_label, tree, to_label, to_entry, to_tree,
504
output_to, reverse=False):
505
"""See InventoryEntry._diff."""
506
from_text = tree.get_file(self.file_id).readlines()
507
if to_entry:
508
to_text = to_tree.get_file(to_entry.file_id).readlines()
509
else:
510
to_text = []
511
if not reverse:
512
text_diff(from_label, from_text,
513
to_label, to_text, output_to)
514
else:
515
text_diff(to_label, to_text,
516
from_label, from_text, output_to)
517
518
def has_text(self):
519
"""See InventoryEntry.has_text."""
520
return True
521
522
def __init__(self, file_id, name, parent_id):
523
super(InventoryFile, self).__init__(file_id, name, parent_id)
524
self.kind = 'file'
525
526
def kind_character(self):
527
"""See InventoryEntry.kind_character."""
528
return ''
529
530
def _put_in_tar(self, item, tree):
531
"""See InventoryEntry._put_in_tar."""
532
item.type = tarfile.REGTYPE
533
fileobj = tree.get_file(self.file_id)
534
item.size = self.text_size
535
if tree.is_executable(self.file_id):
536
item.mode = 0755
537
else:
538
item.mode = 0644
539
return fileobj
540
541
def _put_on_disk(self, fullpath, tree):
542
"""See InventoryEntry._put_on_disk."""
543
pumpfile(tree.get_file(self.file_id), file(fullpath, 'wb'))
544
if tree.is_executable(self.file_id):
545
os.chmod(fullpath, 0755)
546
547
def _read_tree_state(self, path, work_tree):
548
"""See InventoryEntry._read_tree_state."""
549
self.text_sha1 = work_tree.get_file_sha1(self.file_id)
550
self.executable = work_tree.is_executable(self.file_id)
551
552
def _snapshot_text(self, file_parents, work_tree, weave_store, transaction):
553
"""See InventoryEntry._snapshot_text."""
554
mutter('storing file {%s} in revision {%s}',
555
self.file_id, self.revision)
556
# special case to avoid diffing on renames or
557
# reparenting
558
if (len(file_parents) == 1
559
and self.text_sha1 == file_parents.values()[0].text_sha1
560
and self.text_size == file_parents.values()[0].text_size):
561
previous_ie = file_parents.values()[0]
562
weave_store.add_identical_text(
563
self.file_id, previous_ie.revision,
564
self.revision, file_parents, transaction)
565
else:
566
new_lines = work_tree.get_file(self.file_id).readlines()
567
self._add_text_to_weave(new_lines, file_parents, weave_store,
568
transaction)
569
self.text_sha1 = sha_strings(new_lines)
570
self.text_size = sum(map(len, new_lines))
571
572
573
def _unchanged(self, previous_ie):
574
"""See InventoryEntry._unchanged."""
575
compatible = super(InventoryFile, self)._unchanged(previous_ie)
576
if self.text_sha1 != previous_ie.text_sha1:
577
compatible = False
578
else:
579
# FIXME: 20050930 probe for the text size when getting sha1
580
# in _read_tree_state
581
self.text_size = previous_ie.text_size
582
if self.executable != previous_ie.executable:
583
compatible = False
584
return compatible
585
586
587
class InventoryLink(InventoryEntry):
588
"""A file in an inventory."""
589
590
__slots__ = ['symlink_target']
591
592
def _check(self, checker, rev_id, tree):
593
"""See InventoryEntry._check"""
594
if self.text_sha1 != None or self.text_size != None or self.text_id != None:
595
raise BzrCheckError('symlink {%s} has text in revision {%s}'
596
% (self.file_id, rev_id))
597
if self.symlink_target == None:
598
raise BzrCheckError('symlink {%s} has no target in revision {%s}'
599
% (self.file_id, rev_id))
600
601
def copy(self):
602
other = InventoryLink(self.file_id, self.name, self.parent_id)
603
other.symlink_target = self.symlink_target
604
other.revision = self.revision
605
return other
606
607
def detect_changes(self, old_entry):
608
"""See InventoryEntry.detect_changes."""
609
# FIXME: which _modified field should we use ? RBC 20051003
610
text_modified = (self.symlink_target != old_entry.symlink_target)
611
if text_modified:
612
mutter(" symlink target changed")
613
meta_modified = False
614
return text_modified, meta_modified
615
616
def _diff(self, text_diff, from_label, tree, to_label, to_entry, to_tree,
617
output_to, reverse=False):
618
"""See InventoryEntry._diff."""
619
from_text = self.symlink_target
620
if to_entry is not None:
621
to_text = to_entry.symlink_target
622
if reverse:
623
temp = from_text
624
from_text = to_text
625
to_text = temp
626
print >>output_to, '=== target changed %r => %r' % (from_text, to_text)
627
else:
628
if not reverse:
629
print >>output_to, '=== target was %r' % self.symlink_target
630
else:
631
print >>output_to, '=== target is %r' % self.symlink_target
632
633
def __init__(self, file_id, name, parent_id):
634
super(InventoryLink, self).__init__(file_id, name, parent_id)
635
self.kind = 'symlink'
636
637
def kind_character(self):
638
"""See InventoryEntry.kind_character."""
639
return ''
640
641
def _put_in_tar(self, item, tree):
642
"""See InventoryEntry._put_in_tar."""
643
iterm.type = tarfile.SYMTYPE
644
fileobj = None
645
item.size = 0
646
item.mode = 0755
647
item.linkname = self.symlink_target
648
return fileobj
649
650
def _put_on_disk(self, fullpath, tree):
651
"""See InventoryEntry._put_on_disk."""
652
try:
653
os.symlink(self.symlink_target, fullpath)
654
except OSError,e:
655
raise BzrError("Failed to create symlink %r -> %r, error: %s" % (fullpath, self.symlink_target, e))
656
657
def _read_tree_state(self, path, work_tree):
658
"""See InventoryEntry._read_tree_state."""
659
self.symlink_target = work_tree.get_symlink_target(self.file_id)
660
661
def _unchanged(self, previous_ie):
662
"""See InventoryEntry._unchanged."""
663
compatible = super(InventoryLink, self)._unchanged(previous_ie)
664
if self.symlink_target != previous_ie.symlink_target:
665
compatible = False
666
return compatible
667
668
669
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):
670
251
"""Inventory of versioned files in a tree.
671
252
672
253
This describes which file_id is present at each point in the tree,
684
265
inserted, other than through the Inventory API.
685
266
686
267
>>> inv = Inventory()
687
>>> inv.add(InventoryFile('123-123', 'hello.c', ROOT_ID))
688
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))
689
272
>>> inv['123-123'].name
690
273
'hello.c'
691
274
700
283
701
284
>>> [x[0] for x in inv.iter_entries()]
702
285
['hello.c']
703
>>> inv = Inventory('TREE_ROOT-12345678-12345678')
704
>>> inv.add(InventoryFile('123-123', 'hello.c', ROOT_ID))
705
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
706
292
"""
707
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):
708
310
"""Create or read an inventory.
709
311
710
312
If a working directory is specified, the inventory is read
714
316
The inventory is created with a default root directory, with
715
317
an id of None.
716
318
"""
717
# We are letting Branch.initialize() create a unique inventory
718
# root id. Rather than generating a random one here.
719
#if root_id is None:
720
# root_id = bzrlib.branch.gen_file_id('TREE_ROOT')
721
self.root = RootEntry(root_id)
319
self.root = RootEntry(ROOT_ID)
722
320
self._byid = {self.root.file_id: self.root}
723
321
724
322
725
def copy(self):
726
other = Inventory(self.root.file_id)
727
# copy recursively so we know directories will be added before
728
# their children. There are more efficient ways than this...
729
for path, entry in self.iter_entries():
730
if entry == self.root:
731
continue
732
other.add(entry.copy())
733
return other
734
735
736
323
def __iter__(self):
737
324
return iter(self._byid)
738
325
757
344
if ie.kind == 'directory':
758
345
for cn, cie in self.iter_entries(from_dir=ie.file_id):
759
346
yield os.path.join(name, cn), cie
760
761
762
def entries(self):
763
"""Return list of (path, ie) for all entries except the root.
764
765
This may be faster than iter_entries.
347
348
349
350
def directories(self):
351
"""Return (path, entry) pairs for all directories.
766
352
"""
767
accum = []
768
def descend(dir_ie, dir_path):
769
kids = dir_ie.children.items()
770
kids.sort()
771
for name, ie in kids:
772
child_path = os.path.join(dir_path, name)
773
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():
774
360
if ie.kind == 'directory':
775
descend(ie, child_path)
776
777
descend(self.root, '')
778
return accum
779
780
781
def directories(self):
782
"""Return (path, entry) pairs for all directories, including the root.
783
"""
784
accum = []
785
def descend(parent_ie, parent_path):
786
accum.append((parent_path, parent_ie))
361
dn.append((ie.name, ie))
362
dn.sort()
787
363
788
kids = [(ie.name, ie) for ie in parent_ie.children.itervalues() if ie.kind == 'directory']
789
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
790
367
791
for name, child_ie in kids:
792
child_path = os.path.join(parent_path, name)
793
descend(child_ie, child_path)
794
descend(self.root, '')
795
return accum
368
for name, ie in descend(self.root):
369
yield name, ie
796
370
797
371
798
372
800
374
"""True if this entry contains a file with given id.
801
375
802
376
>>> inv = Inventory()
803
>>> inv.add(InventoryFile('123', 'foo.c', ROOT_ID))
804
InventoryFile('123', 'foo.c', parent_id='TREE_ROOT')
377
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
805
378
>>> '123' in inv
806
379
True
807
380
>>> '456' in inv
814
387
"""Return the entry for given file_id.
815
388
816
389
>>> inv = Inventory()
817
>>> inv.add(InventoryFile('123123', 'hello.c', ROOT_ID))
818
InventoryFile('123123', 'hello.c', parent_id='TREE_ROOT')
390
>>> inv.add(InventoryEntry('123123', 'hello.c', 'file', ROOT_ID))
819
391
>>> inv['123123'].name
820
392
'hello.c'
821
393
"""
394
if file_id == None:
395
raise BzrError("can't look up file_id None")
396
822
397
try:
823
398
return self._byid[file_id]
824
399
except KeyError:
825
if file_id == None:
826
raise BzrError("can't look up file_id None")
827
else:
828
raise BzrError("file_id {%s} not in inventory" % file_id)
829
830
831
def get_file_kind(self, file_id):
832
return self._byid[file_id].kind
400
raise BzrError("file_id {%s} not in inventory" % file_id)
401
833
402
834
403
def get_child(self, parent_id, filename):
835
404
return self[parent_id].children.get(filename)
839
408
"""Add entry to inventory.
840
409
841
410
To add a file to a branch ready to be committed, use Branch.add,
842
which calls this.
843
844
Returns the new entry object.
845
"""
411
which calls this."""
846
412
if entry.file_id in self._byid:
847
raise BzrError("inventory already contains entry with id {%s}" % entry.file_id)
848
849
if entry.parent_id == ROOT_ID or entry.parent_id is None:
850
entry.parent_id = self.root.file_id
413
bailout("inventory already contains entry with id {%s}" % entry.file_id)
851
414
852
415
try:
853
416
parent = self._byid[entry.parent_id]
854
417
except KeyError:
855
raise BzrError("parent_id {%s} not in inventory" % entry.parent_id)
418
bailout("parent_id {%s} not in inventory" % entry.parent_id)
856
419
857
420
if parent.children.has_key(entry.name):
858
raise BzrError("%s is already versioned" %
421
bailout("%s is already versioned" %
859
422
appendpath(self.id2path(parent.file_id), entry.name))
860
423
861
424
self._byid[entry.file_id] = entry
862
425
parent.children[entry.name] = entry
863
return entry
864
426
865
427
866
428
def add_path(self, relpath, kind, file_id=None):
867
429
"""Add entry from a path.
868
430
869
The immediate parent must already be versioned.
870
871
Returns the new entry object."""
872
from bzrlib.branch import gen_file_id
873
431
The immediate parent must already be versioned"""
874
432
parts = bzrlib.osutils.splitpath(relpath)
875
433
if len(parts) == 0:
876
raise BzrError("cannot re-add root of inventory")
434
bailout("cannot re-add root of inventory")
877
435
878
436
if file_id == None:
879
file_id = gen_file_id(relpath)
880
881
parent_path = parts[:-1]
882
parent_id = self.path2id(parent_path)
883
if parent_id == None:
884
raise NotVersionedError(parent_path)
885
886
if kind == 'directory':
887
ie = InventoryDirectory(file_id, parts[-1], parent_id)
888
elif kind == 'file':
889
ie = InventoryFile(file_id, parts[-1], parent_id)
890
elif kind == 'symlink':
891
ie = InventoryLink(file_id, parts[-1], parent_id)
892
else:
893
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)
894
443
return self.add(ie)
895
444
896
445
898
447
"""Remove entry by id.
899
448
900
449
>>> inv = Inventory()
901
>>> inv.add(InventoryFile('123', 'foo.c', ROOT_ID))
902
InventoryFile('123', 'foo.c', parent_id='TREE_ROOT')
450
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
903
451
>>> '123' in inv
904
452
True
905
453
>>> del inv['123']
921
469
del self[ie.parent_id].children[ie.name]
922
470
923
471
924
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):
925
505
"""Compare two sets by comparing their contents.
926
506
927
507
>>> i1 = Inventory()
928
508
>>> i2 = Inventory()
929
509
>>> i1 == i2
930
510
True
931
>>> i1.add(InventoryFile('123', 'foo', ROOT_ID))
932
InventoryFile('123', 'foo', parent_id='TREE_ROOT')
511
>>> i1.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
933
512
>>> i1 == i2
934
513
False
935
>>> i2.add(InventoryFile('123', 'foo', ROOT_ID))
936
InventoryFile('123', 'foo', parent_id='TREE_ROOT')
514
>>> i2.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
937
515
>>> i1 == i2
938
516
True
939
517
"""
518
if self is other:
519
return 0
520
940
521
if not isinstance(other, Inventory):
941
522
return NotImplemented
942
523
943
if len(self._byid) != len(other._byid):
944
# shortcut: obviously not the same
945
return False
946
947
return self._byid == other._byid
948
949
950
def __ne__(self, other):
951
return not self.__eq__(other)
952
953
954
def __hash__(self):
955
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
956
532
957
533
958
534
def get_idpath(self, file_id):
968
544
try:
969
545
ie = self._byid[file_id]
970
546
except KeyError:
971
raise BzrError("file_id {%s} not found in inventory" % file_id)
547
bailout("file_id {%s} not found in inventory" % file_id)
972
548
p.insert(0, ie.file_id)
973
549
file_id = ie.parent_id
974
550
return p
975
551
976
552
977
553
def id2path(self, file_id):
978
"""Return as a list the path to file_id.
979
980
>>> i = Inventory()
981
>>> e = i.add(InventoryDirectory('src-id', 'src', ROOT_ID))
982
>>> e = i.add(InventoryFile('foo-id', 'foo.c', parent_id='src-id'))
983
>>> print i.id2path('foo-id').replace(os.sep, '/')
984
src/foo.c
985
"""
554
"""Return as a list the path to file_id."""
555
986
556
# get all names, skipping root
987
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:]]
988
558
return os.sep.join(p)
989
559
990
560
1034
604
1035
605
This does not move the working file."""
1036
606
if not is_valid_name(new_name):
1037
raise BzrError("not an acceptable filename: %r" % new_name)
607
bailout("not an acceptable filename: %r" % new_name)
1038
608
1039
609
new_parent = self._byid[new_parent_id]
1040
610
if new_name in new_parent.children:
1041
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)))
1042
612
1043
613
new_parent_idpath = self.get_idpath(new_parent_id)
1044
614
if file_id in new_parent_idpath:
1045
raise BzrError("cannot move directory %r into a subdirectory of itself, %r"
615
bailout("cannot move directory %r into a subdirectory of itself, %r"
1046
616
% (self.id2path(file_id), self.id2path(new_parent_id)))
1047
617
1048
618
file_ie = self._byid[file_id]
1059
629
1060
630
1061
631
1062
_NAME_RE = None
632
_NAME_RE = re.compile(r'^[^/\\]+$')
1063
633
1064
634
def is_valid_name(name):
1065
global _NAME_RE
1066
if _NAME_RE == None:
1067
_NAME_RE = re.compile(r'^[^/\\]+$')
1068
1069
635
return bool(_NAME_RE.match(name))
Loggerhead is a web-based interface for Breezy
Version: 2.0.1