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
25
from bzrlib.xml_serializer import SubElement, Element, Serializer
26
from bzrlib.inventory import ROOT_ID, Inventory, InventoryEntry
27
from bzrlib.revision import Revision
28
from bzrlib.errors import BzrError
35
"'":"'", # FIXME: overkill
42
def _ensure_utf8_re():
43
"""Make sure the _utf8_re and _unicode_re regexes have been compiled."""
44
global _utf8_re, _unicode_re
46
_utf8_re = re.compile('[&<>\'\"]|[\x80-\xff]+')
47
if _unicode_re is None:
48
_unicode_re = re.compile(u'[&<>\'\"\u0080-\uffff]')
51
def _unicode_escape_replace(match, _map=_xml_escape_map):
52
"""Replace a string of non-ascii, non XML safe characters with their escape
54
This will escape both Standard XML escapes, like <>"', etc.
55
As well as escaping non ascii characters, because ElementTree did.
56
This helps us remain compatible to older versions of bzr. We may change
57
our policy in the future, though.
59
# jam 20060816 Benchmarks show that try/KeyError is faster if you
60
# expect the entity to rarely miss. There is about a 10% difference
61
# in overall time. But if you miss frequently, then if None is much
62
# faster. For our use case, we *rarely* have a revision id, file id
63
# or path name that is unicode. So use try/KeyError.
65
return _map[match.group()]
67
return "&#%d;" % ord(match.group())
70
def _utf8_escape_replace(match, _map=_xml_escape_map):
71
"""Escape utf8 characters into XML safe ones.
73
This uses 2 tricks. It is either escaping "standard" characters, like "&<>,
74
or it is handling characters with the high-bit set. For ascii characters,
75
we just lookup the replacement in the dictionary. For everything else, we
76
decode back into Unicode, and then use the XML escape code.
79
return _map[match.group()]
81
return ''.join('&#%d;' % ord(uni_chr)
82
for uni_chr in match.group().decode('utf8'))
87
def _encode_and_escape(unicode_or_utf8_str, _map=_to_escaped_map):
88
"""Encode the string into utf8, and escape invalid XML characters"""
89
# We frequently get entities we have not seen before, so it is better
90
# to check if None, rather than try/KeyError
91
text = _map.get(unicode_or_utf8_str)
93
if unicode_or_utf8_str.__class__ == unicode:
94
# The alternative policy is to do a regular UTF8 encoding
95
# and then escape only XML meta characters.
96
# Performance is equivalent once you use cache_utf8. *However*
97
# this makes the serialized texts incompatible with old versions
98
# of bzr. So no net gain. (Perhaps the read code would handle utf8
99
# better than entity escapes, but cElementTree seems to do just fine
101
text = str(_unicode_re.sub(_unicode_escape_replace,
102
unicode_or_utf8_str)) + '"'
104
# Plain strings are considered to already be in utf-8 so we do a
105
# slightly different method for escaping.
106
text = _utf8_re.sub(_utf8_escape_replace,
107
unicode_or_utf8_str) + '"'
108
_map[unicode_or_utf8_str] = text
112
def _get_utf8_or_ascii(a_str,
113
_encode_utf8=cache_utf8.encode,
114
_get_cached_ascii=cache_utf8.get_cached_ascii):
115
"""Return a cached version of the string.
117
cElementTree will return a plain string if the XML is plain ascii. It only
118
returns Unicode when it needs to. We want to work in utf-8 strings. So if
119
cElementTree returns a plain string, we can just return the cached version.
120
If it is Unicode, then we need to encode it.
122
:param a_str: An 8-bit string or Unicode as returned by
123
cElementTree.Element.get()
124
:return: A utf-8 encoded 8-bit string.
126
# This is fairly optimized because we know what cElementTree does, this is
127
# not meant as a generic function for all cases. Because it is possible for
128
# an 8-bit string to not be ascii or valid utf8.
129
if a_str.__class__ == unicode:
130
return _encode_utf8(a_str)
132
return _get_cached_ascii(a_str)
136
"""Clean out the unicode => escaped map"""
137
_to_escaped_map.clear()
140
class Serializer_v5(Serializer):
141
"""Version 5 serializer
143
Packs objects into XML and vice versa.
148
support_altered_by_hack = True
149
# This format supports the altered-by hack that reads file ids directly out
150
# of the versionedfile, without doing XML parsing.
152
supported_kinds = set(['file', 'directory', 'symlink'])
155
def write_inventory_to_string(self, inv):
156
"""Just call write_inventory with a StringIO and return the value"""
157
sio = cStringIO.StringIO()
158
self.write_inventory(inv, sio)
159
return sio.getvalue()
161
def write_inventory(self, inv, f):
162
"""Write inventory to a file.
164
:param inv: the inventory to write.
165
:param f: the file to write.
169
append = output.append
170
self._append_inventory_root(append, inv)
171
entries = inv.iter_entries()
173
root_path, root_ie = entries.next()
174
for path, ie in entries:
175
self._append_entry(append, ie)
176
append('</inventory>\n')
178
# Just to keep the cache from growing without bounds
179
# but we may actually not want to do clear the cache
182
def _append_inventory_root(self, append, inv):
183
"""Append the inventory root to output."""
185
if inv.root.file_id not in (None, ROOT_ID):
187
append(_encode_and_escape(inv.root.file_id))
188
append(' format="5"')
189
if inv.revision_id is not None:
190
append(' revision_id="')
191
append(_encode_and_escape(inv.revision_id))
194
def _append_entry(self, append, ie):
195
"""Convert InventoryEntry to XML element and append to output."""
196
# TODO: should just be a plain assertion
197
if ie.kind not in self.supported_kinds:
198
raise errors.UnsupportedInventoryKind(ie.kind)
203
append(' executable="yes"')
205
append(_encode_and_escape(ie.file_id))
207
append(_encode_and_escape(ie.name))
208
if self._parent_condition(ie):
209
assert isinstance(ie.parent_id, basestring)
210
append(' parent_id="')
211
append(_encode_and_escape(ie.parent_id))
212
if ie.revision is not None:
213
append(' revision="')
214
append(_encode_and_escape(ie.revision))
215
if ie.symlink_target is not None:
216
append(' symlink_target="')
217
append(_encode_and_escape(ie.symlink_target))
218
if ie.text_sha1 is not None:
219
append(' text_sha1="')
222
if ie.text_size is not None:
223
append(' text_size="%d"' % ie.text_size)
224
if getattr(ie, 'reference_revision', None) is not None:
225
append(' reference_revision="')
226
append(_encode_and_escape(ie.reference_revision))
230
def _parent_condition(self, ie):
231
return ie.parent_id != ROOT_ID
233
def _pack_revision(self, rev):
234
"""Revision object -> xml tree"""
235
# For the XML format, we need to write them as Unicode rather than as
236
# utf-8 strings. So that cElementTree can handle properly escaping
238
decode_utf8 = cache_utf8.decode
239
revision_id = rev.revision_id
240
if isinstance(revision_id, str):
241
revision_id = decode_utf8(revision_id)
242
root = Element('revision',
243
committer = rev.committer,
244
timestamp = '%.3f' % rev.timestamp,
245
revision_id = revision_id,
246
inventory_sha1 = rev.inventory_sha1,
249
if rev.timezone is not None:
250
root.set('timezone', str(rev.timezone))
252
msg = SubElement(root, 'message')
253
msg.text = rev.message
256
pelts = SubElement(root, 'parents')
257
pelts.tail = pelts.text = '\n'
258
for parent_id in rev.parent_ids:
259
assert isinstance(parent_id, basestring)
260
p = SubElement(pelts, 'revision_ref')
262
if isinstance(parent_id, str):
263
parent_id = decode_utf8(parent_id)
264
p.set('revision_id', parent_id)
266
self._pack_revision_properties(rev, root)
269
def _pack_revision_properties(self, rev, under_element):
270
top_elt = SubElement(under_element, 'properties')
271
for prop_name, prop_value in sorted(rev.properties.items()):
272
assert isinstance(prop_name, basestring)
273
assert isinstance(prop_value, basestring)
274
prop_elt = SubElement(top_elt, 'property')
275
prop_elt.set('name', prop_name)
276
prop_elt.text = prop_value
280
def _unpack_inventory(self, elt):
281
"""Construct from XML Element
283
assert elt.tag == 'inventory'
284
root_id = elt.get('file_id') or ROOT_ID
285
root_id = _get_utf8_or_ascii(root_id)
287
format = elt.get('format')
288
if format is not None:
290
raise BzrError("invalid format version %r on inventory"
292
revision_id = elt.get('revision_id')
293
if revision_id is not None:
294
revision_id = cache_utf8.encode(revision_id)
295
inv = Inventory(root_id, revision_id=revision_id)
297
ie = self._unpack_entry(e)
298
if ie.parent_id is None:
299
ie.parent_id = root_id
303
def _unpack_entry(self, elt):
305
if not InventoryEntry.versionable_kind(kind):
306
raise AssertionError('unsupported entry kind %s' % kind)
308
get_cached = _get_utf8_or_ascii
310
parent_id = elt.get('parent_id')
311
if parent_id is not None:
312
parent_id = get_cached(parent_id)
313
file_id = get_cached(elt.get('file_id'))
315
if kind == 'directory':
316
ie = inventory.InventoryDirectory(file_id,
320
ie = inventory.InventoryFile(file_id,
323
ie.text_sha1 = elt.get('text_sha1')
324
if elt.get('executable') == 'yes':
326
v = elt.get('text_size')
327
ie.text_size = v and int(v)
328
elif kind == 'symlink':
329
ie = inventory.InventoryLink(file_id,
332
ie.symlink_target = elt.get('symlink_target')
334
raise errors.UnsupportedInventoryKind(kind)
335
revision = elt.get('revision')
336
if revision is not None:
337
revision = get_cached(revision)
338
ie.revision = revision
342
def _unpack_revision(self, elt):
343
"""XML Element -> Revision object"""
344
assert elt.tag == 'revision'
345
format = elt.get('format')
346
if format is not None:
348
raise BzrError("invalid format version %r on inventory"
350
get_cached = _get_utf8_or_ascii
351
rev = Revision(committer = elt.get('committer'),
352
timestamp = float(elt.get('timestamp')),
353
revision_id = get_cached(elt.get('revision_id')),
354
inventory_sha1 = elt.get('inventory_sha1')
356
parents = elt.find('parents') or []
358
assert p.tag == 'revision_ref', \
359
"bad parent node tag %r" % p.tag
360
rev.parent_ids.append(get_cached(p.get('revision_id')))
361
self._unpack_revision_properties(elt, rev)
362
v = elt.get('timezone')
366
rev.timezone = int(v)
367
rev.message = elt.findtext('message') # text of <message>
370
def _unpack_revision_properties(self, elt, rev):
371
"""Unpack properties onto a revision."""
372
props_elt = elt.find('properties')
373
assert len(rev.properties) == 0
376
for prop_elt in props_elt:
377
assert prop_elt.tag == 'property', \
378
"bad tag under properties list: %r" % prop_elt.tag
379
name = prop_elt.get('name')
380
value = prop_elt.text
381
# If a property had an empty value ('') cElementTree reads
382
# that back as None, convert it back to '', so that all
383
# properties have string values
386
assert name not in rev.properties, \
387
"repeated property %r" % name
388
rev.properties[name] = value
391
serializer_v5 = Serializer_v5()
added
removed
bzrlib/xml5.py
