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
87
def _get_utf8_or_ascii(a_str,
88
_encode_utf8=cache_utf8.encode,
89
_get_cached_ascii=cache_utf8.get_cached_ascii):
90
"""Return a cached version of the string.
92
cElementTree will return a plain string if the XML is plain ascii. It only
93
returns Unicode when it needs to. We want to work in utf-8 strings. So if
94
cElementTree returns a plain string, we can just return the cached version.
95
If it is Unicode, then we need to encode it.
97
:param a_str: An 8-bit string or Unicode as returned by
98
cElementTree.Element.get()
99
:return: A utf-8 encoded 8-bit string.
101
# This is fairly optimized because we know what cElementTree does, this is
102
# not meant as a generic function for all cases. Because it is possible for
103
# an 8-bit string to not be ascii or valid utf8.
104
if a_str.__class__ == unicode:
105
return _encode_utf8(a_str)
107
return _get_cached_ascii(a_str)
111
"""Clean out the unicode => escaped map"""
112
_unicode_to_escaped_map.clear()
115
class Serializer_v5(Serializer):
116
"""Version 5 serializer
118
Packs objects into XML and vice versa.
123
support_altered_by_hack = True
124
# This format supports the altered-by hack that reads file ids directly out
125
# of the versionedfile, without doing XML parsing.
127
def write_inventory_to_string(self, inv):
128
"""Just call write_inventory with a StringIO and return the value"""
129
sio = cStringIO.StringIO()
130
self.write_inventory(inv, sio)
131
return sio.getvalue()
133
def write_inventory(self, inv, f):
134
"""Write inventory to a file.
136
:param inv: the inventory to write.
137
:param f: the file to write.
141
append = output.append
142
self._append_inventory_root(append, inv)
143
entries = inv.iter_entries()
145
root_path, root_ie = entries.next()
146
for path, ie in entries:
147
self._append_entry(append, ie)
148
append('</inventory>\n')
150
# Just to keep the cache from growing without bounds
151
# but we may actually not want to do clear the cache
154
def _append_inventory_root(self, append, inv):
155
"""Append the inventory root to output."""
157
if inv.root.file_id not in (None, ROOT_ID):
159
append(_encode_and_escape(inv.root.file_id))
160
append(' format="5"')
161
if inv.revision_id is not None:
162
append(' revision_id="')
163
append(_encode_and_escape(inv.revision_id))
166
def _append_entry(self, append, ie):
167
"""Convert InventoryEntry to XML element and append to output."""
168
# TODO: should just be a plain assertion
169
assert InventoryEntry.versionable_kind(ie.kind), \
170
'unsupported entry kind %s' % ie.kind
175
append(' executable="yes"')
177
append(_encode_and_escape(ie.file_id))
179
append(_encode_and_escape(ie.name))
180
if self._parent_condition(ie):
181
assert isinstance(ie.parent_id, basestring)
182
append(' parent_id="')
183
append(_encode_and_escape(ie.parent_id))
184
if ie.revision is not None:
185
append(' revision="')
186
append(_encode_and_escape(ie.revision))
187
if ie.symlink_target is not None:
188
append(' symlink_target="')
189
append(_encode_and_escape(ie.symlink_target))
190
if ie.text_sha1 is not None:
191
append(' text_sha1="')
194
if ie.text_size is not None:
195
append(' text_size="%d"' % ie.text_size)
199
def _parent_condition(self, ie):
200
return ie.parent_id != ROOT_ID
202
def _pack_revision(self, rev):
203
"""Revision object -> xml tree"""
204
# For the XML format, we need to write them as Unicode rather than as
205
# utf-8 strings. So that cElementTree can handle properly escaping
207
decode_utf8 = cache_utf8.decode
208
revision_id = rev.revision_id
209
if isinstance(revision_id, str):
210
revision_id = decode_utf8(revision_id)
211
root = Element('revision',
212
committer = rev.committer,
213
timestamp = '%.3f' % rev.timestamp,
214
revision_id = revision_id,
215
inventory_sha1 = rev.inventory_sha1,
218
if rev.timezone is not None:
219
root.set('timezone', str(rev.timezone))
221
msg = SubElement(root, 'message')
222
msg.text = rev.message
225
pelts = SubElement(root, 'parents')
226
pelts.tail = pelts.text = '\n'
227
for parent_id in rev.parent_ids:
228
assert isinstance(parent_id, basestring)
229
p = SubElement(pelts, 'revision_ref')
231
if isinstance(parent_id, str):
232
parent_id = decode_utf8(parent_id)
233
p.set('revision_id', parent_id)
235
self._pack_revision_properties(rev, root)
238
def _pack_revision_properties(self, rev, under_element):
239
top_elt = SubElement(under_element, 'properties')
240
for prop_name, prop_value in sorted(rev.properties.items()):
241
assert isinstance(prop_name, basestring)
242
assert isinstance(prop_value, basestring)
243
prop_elt = SubElement(top_elt, 'property')
244
prop_elt.set('name', prop_name)
245
prop_elt.text = prop_value
249
def _unpack_inventory(self, elt):
250
"""Construct from XML Element
252
assert elt.tag == 'inventory'
253
root_id = elt.get('file_id') or ROOT_ID
254
format = elt.get('format')
255
if format is not None:
257
raise BzrError("invalid format version %r on inventory"
259
revision_id = elt.get('revision_id')
260
if revision_id is not None:
261
revision_id = cache_utf8.encode(revision_id)
262
inv = Inventory(root_id, revision_id=revision_id)
264
ie = self._unpack_entry(e)
265
if ie.parent_id == ROOT_ID:
266
ie.parent_id = root_id
270
def _unpack_entry(self, elt, none_parents=False):
272
if not InventoryEntry.versionable_kind(kind):
273
raise AssertionError('unsupported entry kind %s' % kind)
275
get_cached = _get_utf8_or_ascii
277
parent_id = elt.get('parent_id')
278
if parent_id is None and not none_parents:
280
# TODO: jam 20060817 At present, caching file ids costs us too
281
# much time. It slows down overall read performances from
282
# approx 500ms to 700ms. And doesn't improve future reads.
283
# it might be because revision ids and file ids are mixing.
284
# Consider caching *just* the file ids, for a limited period
286
#parent_id = get_cached(parent_id)
287
#file_id = get_cached(elt.get('file_id'))
288
file_id = elt.get('file_id')
290
if kind == 'directory':
291
ie = inventory.InventoryDirectory(file_id,
295
ie = inventory.InventoryFile(file_id,
298
ie.text_sha1 = elt.get('text_sha1')
299
if elt.get('executable') == 'yes':
301
v = elt.get('text_size')
302
ie.text_size = v and int(v)
303
elif kind == 'symlink':
304
ie = inventory.InventoryLink(file_id,
307
ie.symlink_target = elt.get('symlink_target')
309
raise BzrError("unknown kind %r" % kind)
310
revision = elt.get('revision')
311
if revision is not None:
312
revision = get_cached(revision)
313
ie.revision = revision
317
def _unpack_revision(self, elt):
318
"""XML Element -> Revision object"""
319
assert elt.tag == 'revision'
320
format = elt.get('format')
321
if format is not None:
323
raise BzrError("invalid format version %r on inventory"
325
get_cached = _get_utf8_or_ascii
326
rev = Revision(committer = elt.get('committer'),
327
timestamp = float(elt.get('timestamp')),
328
revision_id = get_cached(elt.get('revision_id')),
329
inventory_sha1 = elt.get('inventory_sha1')
331
parents = elt.find('parents') or []
333
assert p.tag == 'revision_ref', \
334
"bad parent node tag %r" % p.tag
335
rev.parent_ids.append(get_cached(p.get('revision_id')))
336
self._unpack_revision_properties(elt, rev)
337
v = elt.get('timezone')
341
rev.timezone = int(v)
342
rev.message = elt.findtext('message') # text of <message>
345
def _unpack_revision_properties(self, elt, rev):
346
"""Unpack properties onto a revision."""
347
props_elt = elt.find('properties')
348
assert len(rev.properties) == 0
351
for prop_elt in props_elt:
352
assert prop_elt.tag == 'property', \
353
"bad tag under properties list: %r" % prop_elt.tag
354
name = prop_elt.get('name')
355
value = prop_elt.text
356
# If a property had an empty value ('') cElementTree reads
357
# that back as None, convert it back to '', so that all
358
# properties have string values
361
assert name not in rev.properties, \
362
"repeated property %r" % name
363
rev.properties[name] = value
366
serializer_v5 = Serializer_v5()
added
removed
bzrlib/xml5.py
