1
# Copyright (C) 2005, 2006 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
24
from bzrlib.xml_serializer import SubElement, Element, Serializer
25
from bzrlib.inventory import ROOT_ID, Inventory, InventoryEntry
26
from bzrlib.revision import Revision
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()
92
class Serializer_v5(Serializer):
93
"""Version 5 serializer
95
Packs objects into XML and vice versa.
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."""
130
if inv.root.file_id not in (None, ROOT_ID):
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
148
append(' executable="yes"')
150
append(_encode_and_escape(ie.file_id))
152
append(_encode_and_escape(ie.name))
153
if self._parent_condition(ie):
154
assert isinstance(ie.parent_id, basestring)
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)
172
def _parent_condition(self, ie):
173
return ie.parent_id != ROOT_ID
175
def _pack_revision(self, rev):
176
"""Revision object -> xml tree"""
177
root = Element('revision',
178
committer = rev.committer,
179
timestamp = '%.9f' % rev.timestamp,
180
revision_id = rev.revision_id,
181
inventory_sha1 = rev.inventory_sha1,
184
if rev.timezone is not None:
185
root.set('timezone', str(rev.timezone))
187
msg = SubElement(root, 'message')
188
msg.text = rev.message
191
pelts = SubElement(root, 'parents')
192
pelts.tail = pelts.text = '\n'
193
for parent_id in rev.parent_ids:
194
assert isinstance(parent_id, basestring)
195
p = SubElement(pelts, 'revision_ref')
197
p.set('revision_id', parent_id)
199
self._pack_revision_properties(rev, root)
202
def _pack_revision_properties(self, rev, under_element):
203
top_elt = SubElement(under_element, 'properties')
204
for prop_name, prop_value in sorted(rev.properties.items()):
205
assert isinstance(prop_name, basestring)
206
assert isinstance(prop_value, basestring)
207
prop_elt = SubElement(top_elt, 'property')
208
prop_elt.set('name', prop_name)
209
prop_elt.text = prop_value
213
def _unpack_inventory(self, elt):
214
"""Construct from XML Element
216
assert elt.tag == 'inventory'
217
root_id = elt.get('file_id') or ROOT_ID
218
format = elt.get('format')
219
if format is not None:
221
raise BzrError("invalid format version %r on inventory"
223
revision_id = elt.get('revision_id')
224
if revision_id is not None:
225
revision_id = cache_utf8.get_cached_unicode(revision_id)
226
inv = Inventory(root_id, revision_id=revision_id)
228
ie = self._unpack_entry(e)
229
if ie.parent_id == ROOT_ID:
230
ie.parent_id = root_id
234
def _unpack_entry(self, elt, none_parents=False):
236
if not InventoryEntry.versionable_kind(kind):
237
raise AssertionError('unsupported entry kind %s' % kind)
239
get_cached = cache_utf8.get_cached_unicode
241
parent_id = elt.get('parent_id')
242
if parent_id == None and not none_parents:
244
# TODO: jam 20060817 At present, caching file ids costs us too
245
# much time. It slows down overall read performances from
246
# approx 500ms to 700ms. And doesn't improve future reads.
247
# it might be because revision ids and file ids are mixing.
248
# Consider caching *just* the file ids, for a limited period
250
#parent_id = get_cached(parent_id)
251
#file_id = get_cached(elt.get('file_id'))
252
file_id = elt.get('file_id')
254
if kind == 'directory':
255
ie = inventory.InventoryDirectory(file_id,
259
ie = inventory.InventoryFile(file_id,
262
ie.text_sha1 = elt.get('text_sha1')
263
if elt.get('executable') == 'yes':
265
v = elt.get('text_size')
266
ie.text_size = v and int(v)
267
elif kind == 'symlink':
268
ie = inventory.InventoryLink(file_id,
271
ie.symlink_target = elt.get('symlink_target')
273
raise BzrError("unknown kind %r" % kind)
274
revision = elt.get('revision')
275
if revision is not None:
276
revision = get_cached(revision)
277
ie.revision = revision
281
def _unpack_revision(self, elt):
282
"""XML Element -> Revision object"""
283
assert elt.tag == 'revision'
284
format = elt.get('format')
285
if format is not None:
287
raise BzrError("invalid format version %r on inventory"
289
get_cached = cache_utf8.get_cached_unicode
290
rev = Revision(committer = elt.get('committer'),
291
timestamp = float(elt.get('timestamp')),
292
revision_id = get_cached(elt.get('revision_id')),
293
inventory_sha1 = elt.get('inventory_sha1')
295
parents = elt.find('parents') or []
297
assert p.tag == 'revision_ref', \
298
"bad parent node tag %r" % p.tag
299
rev.parent_ids.append(get_cached(p.get('revision_id')))
300
self._unpack_revision_properties(elt, rev)
301
v = elt.get('timezone')
305
rev.timezone = int(v)
306
rev.message = elt.findtext('message') # text of <message>
309
def _unpack_revision_properties(self, elt, rev):
310
"""Unpack properties onto a revision."""
311
props_elt = elt.find('properties')
312
assert len(rev.properties) == 0
315
for prop_elt in props_elt:
316
assert prop_elt.tag == 'property', \
317
"bad tag under properties list: %r" % prop_elt.tag
318
name = prop_elt.get('name')
319
value = prop_elt.text
320
# If a property had an empty value ('') cElementTree reads
321
# that back as None, convert it back to '', so that all
322
# properties have string values
325
assert name not in rev.properties, \
326
"repeated property %r" % name
327
rev.properties[name] = value
330
serializer_v5 = Serializer_v5()
added
removed
bzrlib/xml5.py
