14
14
# along with this program; if not, write to the Free Software
15
15
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
17
"""Inventories map files to their name in a revision."""
19
# TODO: Maybe store inventory_id in the file? Not really needed.
21
__author__ = "Martin Pool <mbp@canonical.com>"
18
24
# This should really be an id randomly assigned when the tree is
19
25
# created, but it's not for now.
23
29
import sys, os.path, types, re
33
from cElementTree import Element, ElementTree, SubElement
35
from elementtree.ElementTree import Element, ElementTree, SubElement
37
from xml import XMLMixin
38
from errors import bailout, BzrError
26
from bzrlib.errors import BzrError, BzrCheckError
28
41
from bzrlib.osutils import uuid, quotefn, splitpath, joinpath, appendpath
29
42
from bzrlib.trace import mutter
31
class InventoryEntry(object):
44
class InventoryEntry(XMLMixin):
32
45
"""Description of a versioned file.
34
47
An InventoryEntry has the following fields, which are also
62
75
>>> i.add(InventoryEntry('2323', 'bye.c', 'file', '123'))
63
76
Traceback (most recent call last):
65
BzrError: inventory already contains entry with id {2323}
78
BzrError: ('inventory already contains entry with id {2323}', [])
66
79
>>> i.add(InventoryEntry('2324', 'bye.c', 'file', '123'))
67
80
>>> i.add(InventoryEntry('2325', 'wibble', 'directory', '123'))
68
81
>>> i.path2id('src/wibble')
84
97
>>> i.id2path('2326')
85
98
'src/wibble/wibble.c'
87
TODO: Maybe also keep the full path of the entry, and the children?
100
:todo: Maybe also keep the full path of the entry, and the children?
88
101
But those depend on its position within a particular inventory, and
89
102
it would be nice not to need to hold the backpointer here.
92
105
# TODO: split InventoryEntry into subclasses for files,
93
106
# directories, etc etc.
98
108
def __init__(self, file_id, name, kind, parent_id, text_id=None):
99
109
"""Create an InventoryEntry
109
119
>>> e = InventoryEntry('123', 'src/hello.c', 'file', ROOT_ID)
110
120
Traceback (most recent call last):
111
BzrCheckError: InventoryEntry name 'src/hello.c' is invalid
121
BzrError: ("InventoryEntry name is not a simple filename: 'src/hello.c'", [])
113
if '/' in name or '\\' in name:
114
raise BzrCheckError('InventoryEntry name %r is invalid' % name)
124
if len(splitpath(name)) != 1:
125
bailout('InventoryEntry name is not a simple filename: %r'
116
128
self.file_id = file_id
130
assert kind in ['file', 'directory']
119
132
self.text_id = text_id
120
133
self.parent_id = parent_id
134
self.text_sha1 = None
135
self.text_size = None
121
136
if kind == 'directory':
122
137
self.children = {}
126
raise BzrError("unhandled entry kind %r" % kind)
139
assert kind == 'file'
130
142
def sorted_children(self):
138
150
self.parent_id, text_id=self.text_id)
139
151
other.text_sha1 = self.text_sha1
140
152
other.text_size = self.text_size
141
# note that children are *not* copied; they're pulled across when
207
215
from_element = classmethod(from_element)
209
def __eq__(self, other):
217
def __cmp__(self, other):
210
220
if not isinstance(other, InventoryEntry):
211
221
return NotImplemented
213
return (self.file_id == other.file_id) \
214
and (self.name == other.name) \
215
and (self.text_sha1 == other.text_sha1) \
216
and (self.text_size == other.text_size) \
217
and (self.text_id == other.text_id) \
218
and (self.parent_id == other.parent_id) \
219
and (self.kind == other.kind)
222
def __ne__(self, other):
223
return not (self == other)
226
raise ValueError('not hashable')
223
return cmp(self.file_id, other.file_id) \
224
or cmp(self.name, other.name) \
225
or cmp(self.text_sha1, other.text_sha1) \
226
or cmp(self.text_size, other.text_size) \
227
or cmp(self.text_id, other.text_id) \
228
or cmp(self.parent_id, other.parent_id) \
229
or cmp(self.kind, other.kind)
235
238
self.parent_id = None
238
def __eq__(self, other):
241
def __cmp__(self, other):
239
244
if not isinstance(other, RootEntry):
240
245
return NotImplemented
242
return (self.file_id == other.file_id) \
243
and (self.children == other.children)
247
class Inventory(object):
246
return cmp(self.file_id, other.file_id) \
247
or cmp(self.children, other.children)
251
class Inventory(XMLMixin):
248
252
"""Inventory of versioned files in a tree.
250
This describes which file_id is present at each point in the tree,
251
and possibly the SHA-1 or other information about the file.
252
Entries can be looked up either by path or by file_id.
254
An Inventory acts like a set of InventoryEntry items. You can
255
also look files up by their file_id or name.
257
May be read from and written to a metadata file in a tree. To
258
manipulate the inventory (for example to add a file), it is read
259
in, modified, and then written back out.
254
261
The inventory represents a typical unix file tree, with
255
262
directories containing files and subdirectories. We never store
262
269
inserted, other than through the Inventory API.
264
271
>>> inv = Inventory()
272
>>> inv.write_xml(sys.stdout)
265
275
>>> inv.add(InventoryEntry('123-123', 'hello.c', 'file', ROOT_ID))
266
276
>>> inv['123-123'].name
278
288
>>> [x[0] for x in inv.iter_entries()]
291
>>> inv.write_xml(sys.stdout)
293
<entry file_id="123-123" kind="file" name="hello.c" />
298
## TODO: Make sure only canonical filenames are stored.
300
## TODO: Do something sensible about the possible collisions on
301
## case-losing filesystems. Perhaps we should just always forbid
304
## TODO: No special cases for root, rather just give it a file id
305
## like everything else.
307
## TODO: Probably change XML serialization to use nesting
281
309
def __init__(self):
282
310
"""Create or read an inventory.
316
344
if ie.kind == 'directory':
317
345
for cn, cie in self.iter_entries(from_dir=ie.file_id):
318
yield os.path.join(name, cn), cie
322
"""Return list of (path, ie) for all entries except the root.
324
This may be faster than iter_entries.
346
yield '/'.join((name, cn)), cie
350
def directories(self):
351
"""Return (path, entry) pairs for all directories.
327
def descend(dir_ie, dir_path):
328
kids = dir_ie.children.items()
330
for name, ie in kids:
331
child_path = os.path.join(dir_path, name)
332
accum.append((child_path, ie))
353
def descend(parent_ie):
354
parent_name = parent_ie.name
355
yield parent_name, parent_ie
357
# directory children in sorted order
359
for ie in parent_ie.children.itervalues():
333
360
if ie.kind == 'directory':
334
descend(ie, child_path)
336
descend(self.root, '')
340
def directories(self):
341
"""Return (path, entry) pairs for all directories, including the root.
344
def descend(parent_ie, parent_path):
345
accum.append((parent_path, parent_ie))
361
dn.append((ie.name, ie))
347
kids = [(ie.name, ie) for ie in parent_ie.children.itervalues() if ie.kind == 'directory']
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
350
for name, child_ie in kids:
351
child_path = os.path.join(parent_path, name)
352
descend(child_ie, child_path)
353
descend(self.root, '')
368
for name, ie in descend(self.root):
376
391
>>> inv['123123'].name
395
raise BzrError("can't look up file_id None")
380
398
return self._byid[file_id]
383
raise BzrError("can't look up file_id None")
385
raise BzrError("file_id {%s} not in inventory" % file_id)
388
def get_file_kind(self, file_id):
389
return self._byid[file_id].kind
400
raise BzrError("file_id {%s} not in inventory" % file_id)
391
403
def get_child(self, parent_id, filename):
392
404
return self[parent_id].children.get(filename)
398
410
To add a file to a branch ready to be committed, use Branch.add,
399
411
which calls this."""
400
412
if entry.file_id in self._byid:
401
raise BzrError("inventory already contains entry with id {%s}" % entry.file_id)
413
bailout("inventory already contains entry with id {%s}" % entry.file_id)
404
416
parent = self._byid[entry.parent_id]
406
raise BzrError("parent_id {%s} not in inventory" % entry.parent_id)
418
bailout("parent_id {%s} not in inventory" % entry.parent_id)
408
420
if parent.children.has_key(entry.name):
409
raise BzrError("%s is already versioned" %
421
bailout("%s is already versioned" %
410
422
appendpath(self.id2path(parent.file_id), entry.name))
412
424
self._byid[entry.file_id] = entry
417
429
"""Add entry from a path.
419
431
The immediate parent must already be versioned"""
420
from bzrlib.errors import NotVersionedError
422
432
parts = bzrlib.osutils.splitpath(relpath)
423
433
if len(parts) == 0:
424
raise BzrError("cannot re-add root of inventory")
434
bailout("cannot re-add root of inventory")
426
436
if file_id == None:
427
from bzrlib.branch import gen_file_id
428
file_id = gen_file_id(relpath)
430
parent_path = parts[:-1]
431
parent_id = self.path2id(parent_path)
432
if parent_id == None:
433
raise NotVersionedError(parent_path)
437
file_id = bzrlib.branch.gen_file_id(relpath)
439
parent_id = self.path2id(parts[:-1])
440
assert parent_id != None
435
441
ie = InventoryEntry(file_id, parts[-1],
436
442
kind=kind, parent_id=parent_id)
437
443
return self.add(ie)
463
469
del self[ie.parent_id].children[ie.name]
473
return Set(self._byid)
466
476
def to_element(self):
467
477
"""Convert to XML Element"""
468
from bzrlib.xml import Element
470
478
e = Element('inventory')
472
480
for path, ie in self.iter_entries():
510
521
if not isinstance(other, Inventory):
511
522
return NotImplemented
513
if len(self._byid) != len(other._byid):
514
# shortcut: obviously not the same
517
return self._byid == other._byid
520
def __ne__(self, other):
521
return not (self == other)
525
raise ValueError('not hashable')
524
if self.id_set() ^ other.id_set():
527
for file_id in self._byid:
528
c = cmp(self[file_id], other[file_id])
529
534
def get_idpath(self, file_id):
600
605
This does not move the working file."""
601
606
if not is_valid_name(new_name):
602
raise BzrError("not an acceptable filename: %r" % new_name)
607
bailout("not an acceptable filename: %r" % new_name)
604
609
new_parent = self._byid[new_parent_id]
605
610
if new_name in new_parent.children:
606
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)))
608
613
new_parent_idpath = self.get_idpath(new_parent_id)
609
614
if file_id in new_parent_idpath:
610
raise BzrError("cannot move directory %r into a subdirectory of itself, %r"
615
bailout("cannot move directory %r into a subdirectory of itself, %r"
611
616
% (self.id2path(file_id), self.id2path(new_parent_id)))
613
618
file_ie = self._byid[file_id]