1
# Copyright (C) 2005, 2006, 2007 Canonical Ltd
3
3
# This program is free software; you can redistribute it and/or modify
4
4
# it under the terms of the GNU General Public License as published by
5
5
# the Free Software Foundation; either version 2 of the License, or
6
6
# (at your option) any later version.
8
8
# This program is distributed in the hope that it will be useful,
9
9
# but WITHOUT ANY WARRANTY; without even the implied warranty of
10
10
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
11
11
# GNU General Public License for more details.
13
13
# You should have received a copy of the GNU General Public License
14
14
# along with this program; if not, write to the Free Software
15
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
18
from bzrlib.xml import ElementTree, SubElement, Element, Serializer
27
19
from bzrlib.inventory import ROOT_ID, Inventory, InventoryEntry
28
from bzrlib.revision import Revision
20
import bzrlib.inventory as inventory
21
from bzrlib.revision import Revision
29
22
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
28
class Serializer_v5(Serializer):
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'])
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."""
36
def _pack_inventory(self, inv):
37
"""Convert to XML Element"""
38
e = Element('inventory',
185
41
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)
42
e.set('file_id', inv.root.file_id)
43
for path, ie in inv.iter_entries():
44
e.append(self._pack_entry(ie))
48
def _pack_entry(self, ie):
49
"""Convert InventoryEntry to XML element"""
50
if not InventoryEntry.versionable_kind(ie.kind):
51
raise AssertionError('unsupported entry kind %s' % ie.kind)
53
e.set('name', ie.name)
54
e.set('file_id', ie.file_id)
56
if ie.text_size != None:
57
e.set('text_size', '%d' % ie.text_size)
59
for f in ['text_sha1', 'revision', 'symlink_target']:
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):
65
e.set('executable', 'yes')
67
# to be conservative, we don't externalize the root pointers
68
# for now, leaving them as null in the xml form. in a future
69
# version it will be implied by nested elements.
70
if ie.parent_id != ROOT_ID:
209
71
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
72
e.set('parent_id', ie.parent_id)
233
79
def _pack_revision(self, rev):
234
80
"""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
81
root = Element('revision',
243
82
committer = rev.committer,
244
timestamp = '%.3f' % rev.timestamp,
245
revision_id = revision_id,
83
timestamp = '%.9f' % rev.timestamp,
84
revision_id = rev.revision_id,
246
85
inventory_sha1 = rev.inventory_sha1,
249
if rev.timezone is not None:
250
89
root.set('timezone', str(rev.timezone))
252
91
msg = SubElement(root, 'message')
257
96
pelts.tail = pelts.text = '\n'
258
97
for parent_id in rev.parent_ids:
259
98
assert isinstance(parent_id, basestring)
260
_mod_revision.check_not_reserved_id(parent_id)
261
99
p = SubElement(pelts, 'revision_ref')
263
if isinstance(parent_id, str):
264
parent_id = decode_utf8(parent_id)
265
101
p.set('revision_id', parent_id)
267
self._pack_revision_properties(rev, root)
270
def _pack_revision_properties(self, rev, under_element):
271
top_elt = SubElement(under_element, 'properties')
272
for prop_name, prop_value in sorted(rev.properties.items()):
273
assert isinstance(prop_name, basestring)
274
assert isinstance(prop_value, basestring)
275
prop_elt = SubElement(top_elt, 'property')
276
prop_elt.set('name', prop_name)
277
prop_elt.text = prop_value
281
105
def _unpack_inventory(self, elt):
282
106
"""Construct from XML Element
284
108
assert elt.tag == 'inventory'
285
109
root_id = elt.get('file_id') or ROOT_ID
286
root_id = _get_utf8_or_ascii(root_id)
288
110
format = elt.get('format')
289
111
if format is not None:
290
112
if format != '5':
291
113
raise BzrError("invalid format version %r on inventory"
293
revision_id = elt.get('revision_id')
294
if revision_id is not None:
295
revision_id = cache_utf8.encode(revision_id)
296
inv = Inventory(root_id, revision_id=revision_id)
115
inv = Inventory(root_id)
298
117
ie = self._unpack_entry(e)
299
if ie.parent_id is None:
118
if ie.parent_id == ROOT_ID:
300
119
ie.parent_id = root_id
304
124
def _unpack_entry(self, elt):
306
126
if not InventoryEntry.versionable_kind(kind):
307
127
raise AssertionError('unsupported entry kind %s' % kind)
309
get_cached = _get_utf8_or_ascii
311
129
parent_id = elt.get('parent_id')
312
if parent_id is not None:
313
parent_id = get_cached(parent_id)
314
file_id = get_cached(elt.get('file_id'))
130
if parent_id == None:
316
133
if kind == 'directory':
317
ie = inventory.InventoryDirectory(file_id,
134
ie = inventory.InventoryDirectory(elt.get('file_id'),
320
137
elif kind == 'file':
321
ie = inventory.InventoryFile(file_id,
138
ie = inventory.InventoryFile(elt.get('file_id'),
324
141
ie.text_sha1 = elt.get('text_sha1')
348
163
if format != '5':
349
164
raise BzrError("invalid format version %r on inventory"
351
get_cached = _get_utf8_or_ascii
352
166
rev = Revision(committer = elt.get('committer'),
353
167
timestamp = float(elt.get('timestamp')),
354
revision_id = get_cached(elt.get('revision_id')),
168
revision_id = elt.get('revision_id'),
355
169
inventory_sha1 = elt.get('inventory_sha1')
357
171
parents = elt.find('parents') or []
358
172
for p in parents:
359
173
assert p.tag == 'revision_ref', \
360
174
"bad parent node tag %r" % p.tag
361
rev.parent_ids.append(get_cached(p.get('revision_id')))
362
self._unpack_revision_properties(elt, rev)
175
rev.parent_ids.append(p.get('revision_id'))
363
177
v = elt.get('timezone')
367
rev.timezone = int(v)
178
rev.timezone = v and int(v)
368
180
rev.message = elt.findtext('message') # text of <message>
371
def _unpack_revision_properties(self, elt, rev):
372
"""Unpack properties onto a revision."""
373
props_elt = elt.find('properties')
374
assert len(rev.properties) == 0
377
for prop_elt in props_elt:
378
assert prop_elt.tag == 'property', \
379
"bad tag under properties list: %r" % prop_elt.tag
380
name = prop_elt.get('name')
381
value = prop_elt.text
382
# If a property had an empty value ('') cElementTree reads
383
# that back as None, convert it back to '', so that all
384
# properties have string values
387
assert name not in rev.properties, \
388
"repeated property %r" % name
389
rev.properties[name] = value
392
184
serializer_v5 = Serializer_v5()
added
removed
bzrlib/xml5.py
