1
# Copyright (C) 2005, 2006 Canonical Ltd
1
3
# This program is free software; you can redistribute it and/or modify
2
4
# it under the terms of the GNU General Public License as published by
3
5
# the Free Software Foundation; either version 2 of the License, or
4
6
# (at your option) any later version.
6
8
# This program is distributed in the hope that it will be useful,
7
9
# but WITHOUT ANY WARRANTY; without even the implied warranty of
8
10
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
9
11
# GNU General Public License for more details.
11
13
# You should have received a copy of the GNU General Public License
12
14
# along with this program; if not, write to the Free Software
13
15
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
16
from bzrlib.xml import ElementTree, SubElement, Element, Serializer
24
from bzrlib.xml_serializer import SubElement, Element, Serializer
17
25
from bzrlib.inventory import ROOT_ID, Inventory, InventoryEntry
18
import bzrlib.inventory as inventory
19
from bzrlib.revision import Revision
26
from bzrlib.revision import Revision
20
27
from bzrlib.errors import BzrError
33
"'":"'", # FIXME: overkill
40
def _ensure_utf8_re():
41
"""Make sure the _utf8_re regex has been compiled"""
43
if _utf8_re is not None:
45
_utf8_re = re.compile(u'[&<>\'\"\u0080-\uffff]')
48
def _utf8_escape_replace(match, _map=_utf8_escape_map):
49
"""Replace a string of non-ascii, non XML safe characters with their escape
51
This will escape both Standard XML escapes, like <>"', etc.
52
As well as escaping non ascii characters, because ElementTree did.
53
This helps us remain compatible to older versions of bzr. We may change
54
our policy in the future, though.
56
# jam 20060816 Benchmarks show that try/KeyError is faster if you
57
# expect the entity to rarely miss. There is about a 10% difference
58
# in overall time. But if you miss frequently, then if None is much
59
# faster. For our use case, we *rarely* have a revision id, file id
60
# or path name that is unicode. So use try/KeyError.
62
return _map[match.group()]
64
return "&#%d;" % ord(match.group())
67
_unicode_to_escaped_map = {}
69
def _encode_and_escape(unicode_str, _map=_unicode_to_escaped_map):
70
"""Encode the string into utf8, and escape invalid XML characters"""
71
# We frequently get entities we have not seen before, so it is better
72
# to check if None, rather than try/KeyError
73
text = _map.get(unicode_str)
75
# The alternative policy is to do a regular UTF8 encoding
76
# and then escape only XML meta characters.
77
# Performance is equivalent once you use cache_utf8. *However*
78
# this makes the serialized texts incompatible with old versions
79
# of bzr. So no net gain. (Perhaps the read code would handle utf8
80
# better than entity escapes, but cElementTree seems to do just fine
82
text = str(_utf8_re.sub(_utf8_escape_replace, unicode_str)) + '"'
83
_map[unicode_str] = text
88
"""Clean out the unicode => escaped map"""
89
_unicode_to_escaped_map.clear()
23
92
class Serializer_v5(Serializer):
24
93
"""Version 5 serializer
31
def _pack_inventory(self, inv):
32
"""Convert to XML Element"""
33
e = Element('inventory',
100
def write_inventory_to_string(self, inv):
101
"""Just call write_inventory with a StringIO and return the value"""
102
sio = cStringIO.StringIO()
103
self.write_inventory(inv, sio)
104
return sio.getvalue()
106
def write_inventory(self, inv, f):
107
"""Write inventory to a file.
109
:param inv: the inventory to write.
110
:param f: the file to write.
114
append = output.append
115
self._append_inventory_root(append, inv)
116
entries = inv.iter_entries()
118
root_path, root_ie = entries.next()
119
for path, ie in entries:
120
self._append_entry(append, ie)
121
append('</inventory>\n')
123
# Just to keep the cache from growing without bounds
124
# but we may actually not want to do clear the cache
127
def _append_inventory_root(self, append, inv):
128
"""Append the inventory root to output."""
36
130
if inv.root.file_id not in (None, ROOT_ID):
37
e.set('file_id', inv.root.file_id)
38
for path, ie in inv.iter_entries():
39
e.append(self._pack_entry(ie))
43
def _pack_entry(self, ie):
44
"""Convert InventoryEntry to XML element"""
45
if not InventoryEntry.versionable_kind(ie.kind):
46
raise AssertionError('unsupported entry kind %s' % ie.kind)
48
e.set('name', ie.name)
49
e.set('file_id', ie.file_id)
51
if ie.text_size != None:
52
e.set('text_size', '%d' % ie.text_size)
54
for f in ['text_sha1', 'revision', 'symlink_target']:
132
append(_encode_and_escape(inv.root.file_id))
133
append(' format="5"')
134
if inv.revision_id is not None:
135
append(' revision_id="')
136
append(_encode_and_escape(inv.revision_id))
139
def _append_entry(self, append, ie):
140
"""Convert InventoryEntry to XML element and append to output."""
141
# TODO: should just be a plain assertion
142
assert InventoryEntry.versionable_kind(ie.kind), \
143
'unsupported entry kind %s' % ie.kind
60
e.set('executable', 'yes')
62
# to be conservative, we don't externalize the root pointers
63
# for now, leaving them as null in the xml form. in a future
64
# version it will be implied by nested elements.
148
append(' executable="yes"')
150
append(_encode_and_escape(ie.file_id))
152
append(_encode_and_escape(ie.name))
65
153
if ie.parent_id != ROOT_ID:
66
154
assert isinstance(ie.parent_id, basestring)
67
e.set('parent_id', ie.parent_id)
155
append(' parent_id="')
156
append(_encode_and_escape(ie.parent_id))
157
if ie.revision is not None:
158
append(' revision="')
159
append(_encode_and_escape(ie.revision))
160
if ie.symlink_target is not None:
161
append(' symlink_target="')
162
append(_encode_and_escape(ie.symlink_target))
163
if ie.text_sha1 is not None:
164
append(' text_sha1="')
167
if ie.text_size is not None:
168
append(' text_size="%d"' % ie.text_size)
74
172
def _pack_revision(self, rev):
75
173
"""Revision object -> xml tree"""
133
231
def _unpack_entry(self, elt):
135
233
if not InventoryEntry.versionable_kind(kind):
136
234
raise AssertionError('unsupported entry kind %s' % kind)
236
get_cached = cache_utf8.get_cached_unicode
138
238
parent_id = elt.get('parent_id')
139
239
if parent_id == None:
140
240
parent_id = ROOT_ID
241
# TODO: jam 20060817 At present, caching file ids costs us too
242
# much time. It slows down overall read performances from
243
# approx 500ms to 700ms. And doesn't improve future reads.
244
# it might be because revision ids and file ids are mixing.
245
# Consider caching *just* the file ids, for a limited period
247
#parent_id = get_cached(parent_id)
248
#file_id = get_cached(elt.get('file_id'))
249
file_id = elt.get('file_id')
142
251
if kind == 'directory':
143
ie = inventory.InventoryDirectory(elt.get('file_id'),
252
ie = inventory.InventoryDirectory(file_id,
146
255
elif kind == 'file':
147
ie = inventory.InventoryFile(elt.get('file_id'),
256
ie = inventory.InventoryFile(file_id,
150
259
ie.text_sha1 = elt.get('text_sha1')
153
262
v = elt.get('text_size')
154
263
ie.text_size = v and int(v)
155
264
elif kind == 'symlink':
156
ie = inventory.InventoryLink(elt.get('file_id'),
265
ie = inventory.InventoryLink(file_id,
159
268
ie.symlink_target = elt.get('symlink_target')
161
270
raise BzrError("unknown kind %r" % kind)
162
ie.revision = elt.get('revision')
271
revision = elt.get('revision')
272
if revision is not None:
273
revision = get_cached(revision)
274
ie.revision = revision
167
278
def _unpack_revision(self, elt):
168
279
"""XML Element -> Revision object"""
169
280
assert elt.tag == 'revision'
172
283
if format != '5':
173
284
raise BzrError("invalid format version %r on inventory"
286
get_cached = cache_utf8.get_cached_unicode
175
287
rev = Revision(committer = elt.get('committer'),
176
288
timestamp = float(elt.get('timestamp')),
177
revision_id = elt.get('revision_id'),
289
revision_id = get_cached(elt.get('revision_id')),
178
290
inventory_sha1 = elt.get('inventory_sha1')
180
292
parents = elt.find('parents') or []
181
293
for p in parents:
182
294
assert p.tag == 'revision_ref', \
183
295
"bad parent node tag %r" % p.tag
184
rev.parent_ids.append(p.get('revision_id'))
296
rev.parent_ids.append(get_cached(p.get('revision_id')))
185
297
self._unpack_revision_properties(elt, rev)
186
298
v = elt.get('timezone')
187
rev.timezone = v and int(v)
302
rev.timezone = int(v)
188
303
rev.message = elt.findtext('message') # text of <message>
192
306
def _unpack_revision_properties(self, elt, rev):
193
307
"""Unpack properties onto a revision."""
194
308
props_elt = elt.find('properties')
198
312
for prop_elt in props_elt:
199
313
assert prop_elt.tag == 'property', \
200
"bad tag under properties list: %r" % p.tag
314
"bad tag under properties list: %r" % prop_elt.tag
201
315
name = prop_elt.get('name')
202
316
value = prop_elt.text
317
# If a property had an empty value ('') cElementTree reads
318
# that back as None, convert it back to '', so that all
319
# properties have string values
203
322
assert name not in rev.properties, \
204
"repeated property %r" % p.name
323
"repeated property %r" % name
205
324
rev.properties[name] = value
added
removed
bzrlib/xml5.py
