1
# Copyright (C) 2005, 2006, 2007 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
revision as _mod_revision,
26
from bzrlib.xml_serializer import SubElement, Element, Serializer
27
from bzrlib.inventory import ROOT_ID, Inventory, InventoryEntry
28
from bzrlib.revision import Revision
29
from bzrlib.errors import BzrError
36
"'":"'", # FIXME: overkill
43
def _ensure_utf8_re():
44
"""Make sure the _utf8_re and _unicode_re regexes have been compiled."""
45
global _utf8_re, _unicode_re
47
_utf8_re = re.compile('[&<>\'\"]|[\x80-\xff]+')
48
if _unicode_re is None:
49
_unicode_re = re.compile(u'[&<>\'\"\u0080-\uffff]')
52
def _unicode_escape_replace(match, _map=_xml_escape_map):
53
"""Replace a string of non-ascii, non XML safe characters with their escape
55
This will escape both Standard XML escapes, like <>"', etc.
56
As well as escaping non ascii characters, because ElementTree did.
57
This helps us remain compatible to older versions of bzr. We may change
58
our policy in the future, though.
60
# jam 20060816 Benchmarks show that try/KeyError is faster if you
61
# expect the entity to rarely miss. There is about a 10% difference
62
# in overall time. But if you miss frequently, then if None is much
63
# faster. For our use case, we *rarely* have a revision id, file id
64
# or path name that is unicode. So use try/KeyError.
66
return _map[match.group()]
68
return "&#%d;" % ord(match.group())
71
def _utf8_escape_replace(match, _map=_xml_escape_map):
72
"""Escape utf8 characters into XML safe ones.
74
This uses 2 tricks. It is either escaping "standard" characters, like "&<>,
75
or it is handling characters with the high-bit set. For ascii characters,
76
we just lookup the replacement in the dictionary. For everything else, we
77
decode back into Unicode, and then use the XML escape code.
80
return _map[match.group()]
82
return ''.join('&#%d;' % ord(uni_chr)
83
for uni_chr in match.group().decode('utf8'))
88
def _encode_and_escape(unicode_or_utf8_str, _map=_to_escaped_map):
89
"""Encode the string into utf8, and escape invalid XML characters"""
90
# We frequently get entities we have not seen before, so it is better
91
# to check if None, rather than try/KeyError
92
text = _map.get(unicode_or_utf8_str)
94
if unicode_or_utf8_str.__class__ == unicode:
95
# The alternative policy is to do a regular UTF8 encoding
96
# and then escape only XML meta characters.
97
# Performance is equivalent once you use cache_utf8. *However*
98
# this makes the serialized texts incompatible with old versions
99
# of bzr. So no net gain. (Perhaps the read code would handle utf8
100
# better than entity escapes, but cElementTree seems to do just fine
102
text = str(_unicode_re.sub(_unicode_escape_replace,
103
unicode_or_utf8_str)) + '"'
105
# Plain strings are considered to already be in utf-8 so we do a
106
# slightly different method for escaping.
107
text = _utf8_re.sub(_utf8_escape_replace,
108
unicode_or_utf8_str) + '"'
109
_map[unicode_or_utf8_str] = text
113
def _get_utf8_or_ascii(a_str,
114
_encode_utf8=cache_utf8.encode,
115
_get_cached_ascii=cache_utf8.get_cached_ascii):
116
"""Return a cached version of the string.
118
cElementTree will return a plain string if the XML is plain ascii. It only
119
returns Unicode when it needs to. We want to work in utf-8 strings. So if
120
cElementTree returns a plain string, we can just return the cached version.
121
If it is Unicode, then we need to encode it.
123
:param a_str: An 8-bit string or Unicode as returned by
124
cElementTree.Element.get()
125
:return: A utf-8 encoded 8-bit string.
127
# This is fairly optimized because we know what cElementTree does, this is
128
# not meant as a generic function for all cases. Because it is possible for
129
# an 8-bit string to not be ascii or valid utf8.
130
if a_str.__class__ == unicode:
131
return _encode_utf8(a_str)
133
return _get_cached_ascii(a_str)
137
"""Clean out the unicode => escaped map"""
138
_to_escaped_map.clear()
141
class Serializer_v5(Serializer):
142
"""Version 5 serializer
144
Packs objects into XML and vice versa.
149
support_altered_by_hack = True
150
# This format supports the altered-by hack that reads file ids directly out
151
# of the versionedfile, without doing XML parsing.
153
supported_kinds = set(['file', 'directory', 'symlink'])
156
def write_inventory_to_string(self, inv):
157
"""Just call write_inventory with a StringIO and return the value"""
158
sio = cStringIO.StringIO()
159
self.write_inventory(inv, sio)
160
return sio.getvalue()
162
def write_inventory(self, inv, f):
163
"""Write inventory to a file.
165
:param inv: the inventory to write.
166
:param f: the file to write.
170
append = output.append
171
self._append_inventory_root(append, inv)
172
entries = inv.iter_entries()
174
root_path, root_ie = entries.next()
175
for path, ie in entries:
176
self._append_entry(append, ie)
177
append('</inventory>\n')
179
# Just to keep the cache from growing without bounds
180
# but we may actually not want to do clear the cache
183
def _append_inventory_root(self, append, inv):
184
"""Append the inventory root to output."""
186
if inv.root.file_id not in (None, ROOT_ID):
188
append(_encode_and_escape(inv.root.file_id))
189
append(' format="5"')
190
if inv.revision_id is not None:
191
append(' revision_id="')
192
append(_encode_and_escape(inv.revision_id))
195
def _append_entry(self, append, ie):
196
"""Convert InventoryEntry to XML element and append to output."""
197
# TODO: should just be a plain assertion
198
if ie.kind not in self.supported_kinds:
199
raise errors.UnsupportedInventoryKind(ie.kind)
204
append(' executable="yes"')
206
append(_encode_and_escape(ie.file_id))
208
append(_encode_and_escape(ie.name))
209
if self._parent_condition(ie):
210
assert isinstance(ie.parent_id, basestring)
211
append(' parent_id="')
212
append(_encode_and_escape(ie.parent_id))
213
if ie.revision is not None:
214
append(' revision="')
215
append(_encode_and_escape(ie.revision))
216
if ie.symlink_target is not None:
217
append(' symlink_target="')
218
append(_encode_and_escape(ie.symlink_target))
219
if ie.text_sha1 is not None:
220
append(' text_sha1="')
223
if ie.text_size is not None:
224
append(' text_size="%d"' % ie.text_size)
225
if getattr(ie, 'reference_revision', None) is not None:
226
append(' reference_revision="')
227
append(_encode_and_escape(ie.reference_revision))
231
def _parent_condition(self, ie):
232
return ie.parent_id != ROOT_ID
234
def _pack_revision(self, rev):
235
"""Revision object -> xml tree"""
236
# For the XML format, we need to write them as Unicode rather than as
237
# utf-8 strings. So that cElementTree can handle properly escaping
239
decode_utf8 = cache_utf8.decode
240
revision_id = rev.revision_id
241
if isinstance(revision_id, str):
242
revision_id = decode_utf8(revision_id)
243
root = Element('revision',
244
committer = rev.committer,
245
timestamp = '%.3f' % rev.timestamp,
246
revision_id = revision_id,
247
inventory_sha1 = rev.inventory_sha1,
250
if rev.timezone is not None:
251
root.set('timezone', str(rev.timezone))
253
msg = SubElement(root, 'message')
254
msg.text = rev.message
257
pelts = SubElement(root, 'parents')
258
pelts.tail = pelts.text = '\n'
259
for parent_id in rev.parent_ids:
260
assert isinstance(parent_id, basestring)
261
_mod_revision.check_not_reserved_id(parent_id)
262
p = SubElement(pelts, 'revision_ref')
264
if isinstance(parent_id, str):
265
parent_id = decode_utf8(parent_id)
266
p.set('revision_id', parent_id)
268
self._pack_revision_properties(rev, root)
271
def _pack_revision_properties(self, rev, under_element):
272
top_elt = SubElement(under_element, 'properties')
273
for prop_name, prop_value in sorted(rev.properties.items()):
274
assert isinstance(prop_name, basestring)
275
assert isinstance(prop_value, basestring)
276
prop_elt = SubElement(top_elt, 'property')
277
prop_elt.set('name', prop_name)
278
prop_elt.text = prop_value
282
def _unpack_inventory(self, elt):
283
"""Construct from XML Element
285
assert elt.tag == 'inventory'
286
root_id = elt.get('file_id') or ROOT_ID
287
root_id = _get_utf8_or_ascii(root_id)
289
format = elt.get('format')
290
if format is not None:
292
raise BzrError("invalid format version %r on inventory"
294
revision_id = elt.get('revision_id')
295
if revision_id is not None:
296
revision_id = cache_utf8.encode(revision_id)
297
inv = Inventory(root_id, revision_id=revision_id)
299
ie = self._unpack_entry(e)
300
if ie.parent_id is None:
301
ie.parent_id = root_id
305
def _unpack_entry(self, elt):
307
if not InventoryEntry.versionable_kind(kind):
308
raise AssertionError('unsupported entry kind %s' % kind)
310
get_cached = _get_utf8_or_ascii
312
parent_id = elt.get('parent_id')
313
if parent_id is not None:
314
parent_id = get_cached(parent_id)
315
file_id = get_cached(elt.get('file_id'))
317
if kind == 'directory':
318
ie = inventory.InventoryDirectory(file_id,
322
ie = inventory.InventoryFile(file_id,
325
ie.text_sha1 = elt.get('text_sha1')
326
if elt.get('executable') == 'yes':
328
v = elt.get('text_size')
329
ie.text_size = v and int(v)
330
elif kind == 'symlink':
331
ie = inventory.InventoryLink(file_id,
334
ie.symlink_target = elt.get('symlink_target')
336
raise errors.UnsupportedInventoryKind(kind)
337
revision = elt.get('revision')
338
if revision is not None:
339
revision = get_cached(revision)
340
ie.revision = revision
344
def _unpack_revision(self, elt):
345
"""XML Element -> Revision object"""
346
assert elt.tag == 'revision'
347
format = elt.get('format')
348
if format is not None:
350
raise BzrError("invalid format version %r on inventory"
352
get_cached = _get_utf8_or_ascii
353
rev = Revision(committer = elt.get('committer'),
354
timestamp = float(elt.get('timestamp')),
355
revision_id = get_cached(elt.get('revision_id')),
356
inventory_sha1 = elt.get('inventory_sha1')
358
parents = elt.find('parents') or []
360
assert p.tag == 'revision_ref', \
361
"bad parent node tag %r" % p.tag
362
rev.parent_ids.append(get_cached(p.get('revision_id')))
363
self._unpack_revision_properties(elt, rev)
364
v = elt.get('timezone')
368
rev.timezone = int(v)
369
rev.message = elt.findtext('message') # text of <message>
372
def _unpack_revision_properties(self, elt, rev):
373
"""Unpack properties onto a revision."""
374
props_elt = elt.find('properties')
375
assert len(rev.properties) == 0
378
for prop_elt in props_elt:
379
assert prop_elt.tag == 'property', \
380
"bad tag under properties list: %r" % prop_elt.tag
381
name = prop_elt.get('name')
382
value = prop_elt.text
383
# If a property had an empty value ('') cElementTree reads
384
# that back as None, convert it back to '', so that all
385
# properties have string values
388
assert name not in rev.properties, \
389
"repeated property %r" % name
390
rev.properties[name] = value
393
serializer_v5 = Serializer_v5()
added
removed
bzrlib/xml5.py
