1
# (C) 2005 Canonical Ltd
3
# This program is free software; you can redistribute it and/or modify
4
# it under the terms of the GNU General Public License as published by
5
# the Free Software Foundation; either version 2 of the License, or
6
# (at your option) any later version.
8
# This program is distributed in the hope that it will be useful,
9
# but WITHOUT ANY WARRANTY; without even the implied warranty of
10
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
11
# GNU General Public License for more details.
13
# You should have received a copy of the GNU General Public License
14
# along with this program; if not, write to the Free Software
15
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
18
# This should really be an id randomly assigned when the tree is
19
# created, but it's not for now.
23
import sys, os.path, types, re
27
from cElementTree import Element, ElementTree, SubElement
29
from elementtree.ElementTree import Element, ElementTree, SubElement
31
from xml import XMLMixin
32
from errors import bailout, BzrError, BzrCheckError
35
from bzrlib.osutils import uuid, quotefn, splitpath, joinpath, appendpath
36
from bzrlib.trace import mutter
38
class InventoryEntry(XMLMixin):
39
"""Description of a versioned file.
41
An InventoryEntry has the following fields, which are also
42
present in the XML inventory-entry element:
45
* *name*: (only the basename within the directory, must not
47
* *kind*: "directory" or "file"
48
* *directory_id*: (if absent/null means the branch root directory)
49
* *text_sha1*: only for files
50
* *text_size*: in bytes, only for files
51
* *text_id*: identifier for the text version, only for files
53
InventoryEntries can also exist inside a WorkingTree
54
inventory, in which case they are not yet bound to a
55
particular revision of the file. In that case the text_sha1,
56
text_size and text_id are absent.
62
>>> i.add(InventoryEntry('123', 'src', 'directory', ROOT_ID))
63
>>> i.add(InventoryEntry('2323', 'hello.c', 'file', parent_id='123'))
64
>>> for j in i.iter_entries():
67
('src', InventoryEntry('123', 'src', kind='directory', parent_id='TREE_ROOT'))
68
('src/hello.c', InventoryEntry('2323', 'hello.c', kind='file', parent_id='123'))
69
>>> i.add(InventoryEntry('2323', 'bye.c', 'file', '123'))
70
Traceback (most recent call last):
72
BzrError: ('inventory already contains entry with id {2323}', [])
73
>>> i.add(InventoryEntry('2324', 'bye.c', 'file', '123'))
74
>>> i.add(InventoryEntry('2325', 'wibble', 'directory', '123'))
75
>>> i.path2id('src/wibble')
79
>>> i.add(InventoryEntry('2326', 'wibble.c', 'file', '2325'))
81
InventoryEntry('2326', 'wibble.c', kind='file', parent_id='2325')
82
>>> for j in i.iter_entries():
84
... assert i.path2id(j[0])
94
TODO: Maybe also keep the full path of the entry, and the children?
95
But those depend on its position within a particular inventory, and
96
it would be nice not to need to hold the backpointer here.
99
# TODO: split InventoryEntry into subclasses for files,
100
# directories, etc etc.
105
def __init__(self, file_id, name, kind, parent_id, text_id=None):
106
"""Create an InventoryEntry
108
The filename must be a single component, relative to the
109
parent directory; it cannot be a whole path or relative name.
111
>>> e = InventoryEntry('123', 'hello.c', 'file', ROOT_ID)
116
>>> e = InventoryEntry('123', 'src/hello.c', 'file', ROOT_ID)
117
Traceback (most recent call last):
118
BzrCheckError: InventoryEntry name 'src/hello.c' is invalid
120
if '/' in name or '\\' in name:
121
raise BzrCheckError('InventoryEntry name %r is invalid' % name)
123
self.file_id = file_id
126
self.text_id = text_id
127
self.parent_id = parent_id
128
if kind == 'directory':
133
raise BzrError("unhandled entry kind %r" % kind)
137
def sorted_children(self):
138
l = self.children.items()
144
other = InventoryEntry(self.file_id, self.name, self.kind,
145
self.parent_id, text_id=self.text_id)
146
other.text_sha1 = self.text_sha1
147
other.text_size = self.text_size
148
# note that children are *not* copied; they're pulled across when
154
return ("%s(%r, %r, kind=%r, parent_id=%r)"
155
% (self.__class__.__name__,
162
def to_element(self):
163
"""Convert to XML element"""
166
e.set('name', self.name)
167
e.set('file_id', self.file_id)
168
e.set('kind', self.kind)
170
if self.text_size != None:
171
e.set('text_size', '%d' % self.text_size)
173
for f in ['text_id', 'text_sha1']:
178
# to be conservative, we don't externalize the root pointers
179
# for now, leaving them as null in the xml form. in a future
180
# version it will be implied by nested elements.
181
if self.parent_id != ROOT_ID:
182
assert isinstance(self.parent_id, basestring)
183
e.set('parent_id', self.parent_id)
190
def from_element(cls, elt):
191
assert elt.tag == 'entry'
193
## original format inventories don't have a parent_id for
194
## nodes in the root directory, but it's cleaner to use one
196
parent_id = elt.get('parent_id')
197
if parent_id == None:
200
self = cls(elt.get('file_id'), elt.get('name'), elt.get('kind'), parent_id)
201
self.text_id = elt.get('text_id')
202
self.text_sha1 = elt.get('text_sha1')
204
## mutter("read inventoryentry: %r" % (elt.attrib))
206
v = elt.get('text_size')
207
self.text_size = v and int(v)
212
from_element = classmethod(from_element)
214
def __cmp__(self, other):
217
if not isinstance(other, InventoryEntry):
218
return NotImplemented
220
return cmp(self.file_id, other.file_id) \
221
or cmp(self.name, other.name) \
222
or cmp(self.text_sha1, other.text_sha1) \
223
or cmp(self.text_size, other.text_size) \
224
or cmp(self.text_id, other.text_id) \
225
or cmp(self.parent_id, other.parent_id) \
226
or cmp(self.kind, other.kind)
230
class RootEntry(InventoryEntry):
231
def __init__(self, file_id):
232
self.file_id = file_id
234
self.kind = 'root_directory'
235
self.parent_id = None
238
def __cmp__(self, other):
241
if not isinstance(other, RootEntry):
242
return NotImplemented
243
return cmp(self.file_id, other.file_id) \
244
or cmp(self.children, other.children)
248
class Inventory(XMLMixin):
249
"""Inventory of versioned files in a tree.
251
This describes which file_id is present at each point in the tree,
252
and possibly the SHA-1 or other information about the file.
253
Entries can be looked up either by path or by file_id.
255
The inventory represents a typical unix file tree, with
256
directories containing files and subdirectories. We never store
257
the full path to a file, because renaming a directory implicitly
258
moves all of its contents. This class internally maintains a
259
lookup tree that allows the children under a directory to be
262
InventoryEntry objects must not be modified after they are
263
inserted, other than through the Inventory API.
265
>>> inv = Inventory()
266
>>> inv.write_xml(sys.stdout)
269
>>> inv.add(InventoryEntry('123-123', 'hello.c', 'file', ROOT_ID))
270
>>> inv['123-123'].name
273
May be treated as an iterator or set to look up file ids:
275
>>> bool(inv.path2id('hello.c'))
280
May also look up by name:
282
>>> [x[0] for x in inv.iter_entries()]
285
>>> inv.write_xml(sys.stdout)
287
<entry file_id="123-123" kind="file" name="hello.c" />
292
"""Create or read an inventory.
294
If a working directory is specified, the inventory is read
295
from there. If the file is specified, read from that. If not,
296
the inventory is created empty.
298
The inventory is created with a default root directory, with
301
self.root = RootEntry(ROOT_ID)
302
self._byid = {self.root.file_id: self.root}
306
return iter(self._byid)
310
"""Returns number of entries."""
311
return len(self._byid)
314
def iter_entries(self, from_dir=None):
315
"""Return (path, entry) pairs, in order by name."""
319
elif isinstance(from_dir, basestring):
320
from_dir = self._byid[from_dir]
322
kids = from_dir.children.items()
324
for name, ie in kids:
326
if ie.kind == 'directory':
327
for cn, cie in self.iter_entries(from_dir=ie.file_id):
328
yield os.path.join(name, cn), cie
332
def directories(self):
333
"""Return (path, entry) pairs for all directories.
335
def descend(parent_ie):
336
parent_name = parent_ie.name
337
yield parent_name, parent_ie
339
# directory children in sorted order
341
for ie in parent_ie.children.itervalues():
342
if ie.kind == 'directory':
343
dn.append((ie.name, ie))
346
for name, child_ie in dn:
347
for sub_name, sub_ie in descend(child_ie):
348
yield appendpath(parent_name, sub_name), sub_ie
350
for name, ie in descend(self.root):
355
def __contains__(self, file_id):
356
"""True if this entry contains a file with given id.
358
>>> inv = Inventory()
359
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
365
return file_id in self._byid
368
def __getitem__(self, file_id):
369
"""Return the entry for given file_id.
371
>>> inv = Inventory()
372
>>> inv.add(InventoryEntry('123123', 'hello.c', 'file', ROOT_ID))
373
>>> inv['123123'].name
377
return self._byid[file_id]
380
raise BzrError("can't look up file_id None")
382
raise BzrError("file_id {%s} not in inventory" % file_id)
385
def get_file_kind(self, file_id):
386
return self._byid[file_id].kind
388
def get_child(self, parent_id, filename):
389
return self[parent_id].children.get(filename)
392
def add(self, entry):
393
"""Add entry to inventory.
395
To add a file to a branch ready to be committed, use Branch.add,
397
if entry.file_id in self._byid:
398
bailout("inventory already contains entry with id {%s}" % entry.file_id)
401
parent = self._byid[entry.parent_id]
403
bailout("parent_id {%s} not in inventory" % entry.parent_id)
405
if parent.children.has_key(entry.name):
406
bailout("%s is already versioned" %
407
appendpath(self.id2path(parent.file_id), entry.name))
409
self._byid[entry.file_id] = entry
410
parent.children[entry.name] = entry
413
def add_path(self, relpath, kind, file_id=None):
414
"""Add entry from a path.
416
The immediate parent must already be versioned"""
417
parts = bzrlib.osutils.splitpath(relpath)
419
bailout("cannot re-add root of inventory")
422
file_id = bzrlib.branch.gen_file_id(relpath)
424
parent_id = self.path2id(parts[:-1])
425
assert parent_id != None
426
ie = InventoryEntry(file_id, parts[-1],
427
kind=kind, parent_id=parent_id)
431
def __delitem__(self, file_id):
432
"""Remove entry by id.
434
>>> inv = Inventory()
435
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
444
assert self[ie.parent_id].children[ie.name] == ie
446
# TODO: Test deleting all children; maybe hoist to a separate
448
if ie.kind == 'directory':
449
for cie in ie.children.values():
450
del self[cie.file_id]
453
del self._byid[file_id]
454
del self[ie.parent_id].children[ie.name]
458
return Set(self._byid)
461
def to_element(self):
462
"""Convert to XML Element"""
463
e = Element('inventory')
465
for path, ie in self.iter_entries():
466
e.append(ie.to_element())
470
def from_element(cls, elt):
471
"""Construct from XML Element
473
>>> inv = Inventory()
474
>>> inv.add(InventoryEntry('foo.c-123981239', 'foo.c', 'file', ROOT_ID))
475
>>> elt = inv.to_element()
476
>>> inv2 = Inventory.from_element(elt)
480
assert elt.tag == 'inventory'
483
o.add(InventoryEntry.from_element(e))
486
from_element = classmethod(from_element)
489
def __cmp__(self, other):
490
"""Compare two sets by comparing their contents.
496
>>> i1.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
499
>>> i2.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
506
if not isinstance(other, Inventory):
507
return NotImplemented
509
if self.id_set() ^ other.id_set():
512
for file_id in self._byid:
513
c = cmp(self[file_id], other[file_id])
519
def get_idpath(self, file_id):
520
"""Return a list of file_ids for the path to an entry.
522
The list contains one element for each directory followed by
523
the id of the file itself. So the length of the returned list
524
is equal to the depth of the file in the tree, counting the
525
root directory as depth 1.
528
while file_id != None:
530
ie = self._byid[file_id]
532
bailout("file_id {%s} not found in inventory" % file_id)
533
p.insert(0, ie.file_id)
534
file_id = ie.parent_id
538
def id2path(self, file_id):
539
"""Return as a list the path to file_id."""
541
# get all names, skipping root
542
p = [self[fid].name for fid in self.get_idpath(file_id)[1:]]
543
return os.sep.join(p)
547
def path2id(self, name):
548
"""Walk down through directories to return entry of last component.
550
names may be either a list of path components, or a single
551
string, in which case it is automatically split.
553
This returns the entry of the last component in the path,
554
which may be either a file or a directory.
556
Returns None iff the path is not found.
558
if isinstance(name, types.StringTypes):
559
name = splitpath(name)
561
mutter("lookup path %r" % name)
566
cie = parent.children[f]
568
assert cie.parent_id == parent.file_id
574
return parent.file_id
577
def has_filename(self, names):
578
return bool(self.path2id(names))
581
def has_id(self, file_id):
582
return self._byid.has_key(file_id)
585
def rename(self, file_id, new_parent_id, new_name):
586
"""Move a file within the inventory.
588
This can change either the name, or the parent, or both.
590
This does not move the working file."""
591
if not is_valid_name(new_name):
592
bailout("not an acceptable filename: %r" % new_name)
594
new_parent = self._byid[new_parent_id]
595
if new_name in new_parent.children:
596
bailout("%r already exists in %r" % (new_name, self.id2path(new_parent_id)))
598
new_parent_idpath = self.get_idpath(new_parent_id)
599
if file_id in new_parent_idpath:
600
bailout("cannot move directory %r into a subdirectory of itself, %r"
601
% (self.id2path(file_id), self.id2path(new_parent_id)))
603
file_ie = self._byid[file_id]
604
old_parent = self._byid[file_ie.parent_id]
606
# TODO: Don't leave things messed up if this fails
608
del old_parent.children[file_ie.name]
609
new_parent.children[new_name] = file_ie
611
file_ie.name = new_name
612
file_ie.parent_id = new_parent_id
617
_NAME_RE = re.compile(r'^[^/\\]+$')
619
def is_valid_name(name):
620
return bool(_NAME_RE.match(name))
added
removed
bzrlib/inventory.py
