1
# Copyright (C) 2005, 2006, 2007, 2008 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_v8(Serializer):
142
"""This serialiser adds rich roots.
144
Its revision format number matches its inventory number.
150
support_altered_by_hack = True
151
# This format supports the altered-by hack that reads file ids directly out
152
# of the versionedfile, without doing XML parsing.
154
supported_kinds = set(['file', 'directory', 'symlink'])
156
revision_format_num = None
158
def _check_revisions(self, inv):
159
"""Extension point for subclasses to check during serialisation.
161
:param inv: An inventory about to be serialised, to be checked.
162
:raises: AssertionError if an error has occured.
164
assert inv.revision_id is not None
165
assert inv.root.revision is not None
167
def write_inventory_to_lines(self, inv):
168
"""Return a list of lines with the encoded inventory."""
169
return self.write_inventory(inv, None)
171
def write_inventory_to_string(self, inv, working=False):
172
"""Just call write_inventory with a StringIO and return the value.
174
:param working: If True skip history data - text_sha1, text_size,
175
reference_revision, symlink_target.
177
sio = cStringIO.StringIO()
178
self.write_inventory(inv, sio, working)
179
return sio.getvalue()
181
def write_inventory(self, inv, f, working=False):
182
"""Write inventory to a file.
184
:param inv: the inventory to write.
185
:param f: the file to write. (May be None if the lines are the desired
187
:param working: If True skip history data - text_sha1, text_size,
188
reference_revision, symlink_target.
189
:return: The inventory as a list of lines.
192
self._check_revisions(inv)
194
append = output.append
195
self._append_inventory_root(append, inv)
196
entries = inv.iter_entries()
198
root_path, root_ie = entries.next()
199
for path, ie in entries:
200
if ie.parent_id != self.root_id:
201
parent_str = ' parent_id="'
202
parent_id = _encode_and_escape(ie.parent_id)
206
if ie.kind == 'file':
208
executable = ' executable="yes"'
212
append('<file%s file_id="%s name="%s%s%s revision="%s '
213
'text_sha1="%s" text_size="%d" />\n' % (
214
executable, _encode_and_escape(ie.file_id),
215
_encode_and_escape(ie.name), parent_str, parent_id,
216
_encode_and_escape(ie.revision), ie.text_sha1,
219
append('<file%s file_id="%s name="%s%s%s />\n' % (
220
executable, _encode_and_escape(ie.file_id),
221
_encode_and_escape(ie.name), parent_str, parent_id))
222
elif ie.kind == 'directory':
224
append('<directory file_id="%s name="%s%s%s revision="%s '
226
_encode_and_escape(ie.file_id),
227
_encode_and_escape(ie.name),
228
parent_str, parent_id,
229
_encode_and_escape(ie.revision)))
231
append('<directory file_id="%s name="%s%s%s />\n' % (
232
_encode_and_escape(ie.file_id),
233
_encode_and_escape(ie.name),
234
parent_str, parent_id))
235
elif ie.kind == 'symlink':
237
append('<symlink file_id="%s name="%s%s%s revision="%s '
238
'symlink_target="%s />\n' % (
239
_encode_and_escape(ie.file_id),
240
_encode_and_escape(ie.name),
241
parent_str, parent_id,
242
_encode_and_escape(ie.revision),
243
_encode_and_escape(ie.symlink_target)))
245
append('<symlink file_id="%s name="%s%s%s />\n' % (
246
_encode_and_escape(ie.file_id),
247
_encode_and_escape(ie.name),
248
parent_str, parent_id))
249
elif ie.kind == 'tree-reference':
250
if ie.kind not in self.supported_kinds:
251
raise errors.UnsupportedInventoryKind(ie.kind)
253
append('<tree-reference file_id="%s name="%s%s%s '
254
'revision="%s reference_revision="%s />\n' % (
255
_encode_and_escape(ie.file_id),
256
_encode_and_escape(ie.name),
257
parent_str, parent_id,
258
_encode_and_escape(ie.revision),
259
_encode_and_escape(ie.reference_revision)))
261
append('<tree-reference file_id="%s name="%s%s%s />\n' % (
262
_encode_and_escape(ie.file_id),
263
_encode_and_escape(ie.name),
264
parent_str, parent_id))
266
raise errors.UnsupportedInventoryKind(ie.kind)
267
append('</inventory>\n')
270
# Just to keep the cache from growing without bounds
271
# but we may actually not want to do clear the cache
275
def _append_inventory_root(self, append, inv):
276
"""Append the inventory root to output."""
277
if inv.revision_id is not None:
278
revid1 = ' revision_id="'
279
revid2 = _encode_and_escape(inv.revision_id)
283
append('<inventory format="%s"%s%s>\n' % (
284
self.format_num, revid1, revid2))
285
append('<directory file_id="%s name="%s revision="%s />\n' % (
286
_encode_and_escape(inv.root.file_id),
287
_encode_and_escape(inv.root.name),
288
_encode_and_escape(inv.root.revision)))
290
def _pack_revision(self, rev):
291
"""Revision object -> xml tree"""
292
# For the XML format, we need to write them as Unicode rather than as
293
# utf-8 strings. So that cElementTree can handle properly escaping
295
decode_utf8 = cache_utf8.decode
296
revision_id = rev.revision_id
297
if isinstance(revision_id, str):
298
revision_id = decode_utf8(revision_id)
299
format_num = self.format_num
300
if self.revision_format_num is not None:
301
format_num = self.revision_format_num
302
root = Element('revision',
303
committer = rev.committer,
304
timestamp = '%.3f' % rev.timestamp,
305
revision_id = revision_id,
306
inventory_sha1 = rev.inventory_sha1,
309
if rev.timezone is not None:
310
root.set('timezone', str(rev.timezone))
312
msg = SubElement(root, 'message')
313
msg.text = rev.message
316
pelts = SubElement(root, 'parents')
317
pelts.tail = pelts.text = '\n'
318
for parent_id in rev.parent_ids:
319
assert isinstance(parent_id, basestring)
320
_mod_revision.check_not_reserved_id(parent_id)
321
p = SubElement(pelts, 'revision_ref')
323
if isinstance(parent_id, str):
324
parent_id = decode_utf8(parent_id)
325
p.set('revision_id', parent_id)
327
self._pack_revision_properties(rev, root)
330
def _pack_revision_properties(self, rev, under_element):
331
top_elt = SubElement(under_element, 'properties')
332
for prop_name, prop_value in sorted(rev.properties.items()):
333
assert isinstance(prop_name, basestring)
334
assert isinstance(prop_value, basestring)
335
prop_elt = SubElement(top_elt, 'property')
336
prop_elt.set('name', prop_name)
337
prop_elt.text = prop_value
341
def _unpack_inventory(self, elt, revision_id=None):
342
"""Construct from XML Element"""
343
if elt.tag != 'inventory':
344
raise errors.UnexpectedInventoryFormat('Root tag is %r' % elt.tag)
345
format = elt.get('format')
346
if format != self.format_num:
347
raise errors.UnexpectedInventoryFormat('Invalid format version %r'
349
revision_id = elt.get('revision_id')
350
if revision_id is not None:
351
revision_id = cache_utf8.encode(revision_id)
352
inv = inventory.Inventory(root_id=None, revision_id=revision_id)
354
ie = self._unpack_entry(e)
356
assert inv.root.revision is not None
359
def _unpack_entry(self, elt):
361
if not InventoryEntry.versionable_kind(kind):
362
raise AssertionError('unsupported entry kind %s' % kind)
364
get_cached = _get_utf8_or_ascii
366
parent_id = elt.get('parent_id')
367
if parent_id is not None:
368
parent_id = get_cached(parent_id)
369
file_id = get_cached(elt.get('file_id'))
371
if kind == 'directory':
372
ie = inventory.InventoryDirectory(file_id,
376
ie = inventory.InventoryFile(file_id,
379
ie.text_sha1 = elt.get('text_sha1')
380
if elt.get('executable') == 'yes':
382
v = elt.get('text_size')
383
ie.text_size = v and int(v)
384
elif kind == 'symlink':
385
ie = inventory.InventoryLink(file_id,
388
ie.symlink_target = elt.get('symlink_target')
390
raise errors.UnsupportedInventoryKind(kind)
391
revision = elt.get('revision')
392
if revision is not None:
393
revision = get_cached(revision)
394
ie.revision = revision
398
def _unpack_revision(self, elt):
399
"""XML Element -> Revision object"""
400
assert elt.tag == 'revision'
401
format = elt.get('format')
402
format_num = self.format_num
403
if self.revision_format_num is not None:
404
format_num = self.revision_format_num
405
if format is not None:
406
if format != format_num:
407
raise BzrError("invalid format version %r on revision"
409
get_cached = _get_utf8_or_ascii
410
rev = Revision(committer = elt.get('committer'),
411
timestamp = float(elt.get('timestamp')),
412
revision_id = get_cached(elt.get('revision_id')),
413
inventory_sha1 = elt.get('inventory_sha1')
415
parents = elt.find('parents') or []
417
assert p.tag == 'revision_ref', \
418
"bad parent node tag %r" % p.tag
419
rev.parent_ids.append(get_cached(p.get('revision_id')))
420
self._unpack_revision_properties(elt, rev)
421
v = elt.get('timezone')
425
rev.timezone = int(v)
426
rev.message = elt.findtext('message') # text of <message>
429
def _unpack_revision_properties(self, elt, rev):
430
"""Unpack properties onto a revision."""
431
props_elt = elt.find('properties')
432
assert len(rev.properties) == 0
435
for prop_elt in props_elt:
436
assert prop_elt.tag == 'property', \
437
"bad tag under properties list: %r" % prop_elt.tag
438
name = prop_elt.get('name')
439
value = prop_elt.text
440
# If a property had an empty value ('') cElementTree reads
441
# that back as None, convert it back to '', so that all
442
# properties have string values
445
assert name not in rev.properties, \
446
"repeated property %r" % name
447
rev.properties[name] = value
450
serializer_v8 = Serializer_v8()
added
removed
bzrlib/xml8.py
