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
from bzrlib.revision import Revision, RevisionReference
29
21
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
27
class Serializer_v5(Serializer):
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'])
157
def write_inventory_to_lines(self, inv):
158
"""Return a list of lines with the encoded inventory."""
159
return self.write_inventory(inv, None)
161
def write_inventory_to_string(self, inv, working=False):
162
"""Just call write_inventory with a StringIO and return the value.
164
:param working: If True skip history data - text_sha1, text_size,
165
reference_revision, symlink_target.
167
sio = cStringIO.StringIO()
168
self.write_inventory(inv, sio, working)
169
return sio.getvalue()
171
def write_inventory(self, inv, f, working=False):
172
"""Write inventory to a file.
174
:param inv: the inventory to write.
175
:param f: the file to write. (May be None if the lines are the desired
177
:param working: If True skip history data - text_sha1, text_size,
178
reference_revision, symlink_target.
179
:return: The inventory as a list of lines.
183
append = output.append
184
self._append_inventory_root(append, inv)
185
entries = inv.iter_entries()
187
root_path, root_ie = entries.next()
188
for path, ie in entries:
189
if ie.parent_id != self.root_id:
190
parent_str = ' parent_id="'
191
parent_id = _encode_and_escape(ie.parent_id)
195
if ie.kind == 'file':
197
executable = ' executable="yes"'
201
append('<file%s file_id="%s name="%s%s%s revision="%s '
202
'text_sha1="%s" text_size="%d" />\n' % (
203
executable, _encode_and_escape(ie.file_id),
204
_encode_and_escape(ie.name), parent_str, parent_id,
205
_encode_and_escape(ie.revision), ie.text_sha1,
208
append('<file%s file_id="%s name="%s%s%s />\n' % (
209
executable, _encode_and_escape(ie.file_id),
210
_encode_and_escape(ie.name), parent_str, parent_id))
211
elif ie.kind == 'directory':
213
append('<directory file_id="%s name="%s%s%s revision="%s '
215
_encode_and_escape(ie.file_id),
216
_encode_and_escape(ie.name),
217
parent_str, parent_id,
218
_encode_and_escape(ie.revision)))
220
append('<directory file_id="%s name="%s%s%s />\n' % (
221
_encode_and_escape(ie.file_id),
222
_encode_and_escape(ie.name),
223
parent_str, parent_id))
224
elif ie.kind == 'symlink':
226
append('<symlink file_id="%s name="%s%s%s revision="%s '
227
'symlink_target="%s />\n' % (
228
_encode_and_escape(ie.file_id),
229
_encode_and_escape(ie.name),
230
parent_str, parent_id,
231
_encode_and_escape(ie.revision),
232
_encode_and_escape(ie.symlink_target)))
234
append('<symlink file_id="%s name="%s%s%s />\n' % (
235
_encode_and_escape(ie.file_id),
236
_encode_and_escape(ie.name),
237
parent_str, parent_id))
238
elif ie.kind == 'tree-reference':
239
if ie.kind not in self.supported_kinds:
240
raise errors.UnsupportedInventoryKind(ie.kind)
242
append('<tree-reference file_id="%s name="%s%s%s '
243
'revision="%s reference_revision="%s />\n' % (
244
_encode_and_escape(ie.file_id),
245
_encode_and_escape(ie.name),
246
parent_str, parent_id,
247
_encode_and_escape(ie.revision),
248
_encode_and_escape(ie.reference_revision)))
250
append('<tree-reference file_id="%s name="%s%s%s />\n' % (
251
_encode_and_escape(ie.file_id),
252
_encode_and_escape(ie.name),
253
parent_str, parent_id))
255
raise errors.UnsupportedInventoryKind(ie.kind)
256
append('</inventory>\n')
259
# Just to keep the cache from growing without bounds
260
# but we may actually not want to do clear the cache
264
def _append_inventory_root(self, append, inv):
265
"""Append the inventory root to output."""
35
def _pack_inventory(self, inv):
36
"""Convert to XML Element"""
37
e = Element('inventory')
266
39
if inv.root.file_id not in (None, ROOT_ID):
267
fileid1 = ' file_id="'
268
fileid2 = _encode_and_escape(inv.root.file_id)
272
if inv.revision_id is not None:
273
revid1 = ' revision_id="'
274
revid2 = _encode_and_escape(inv.revision_id)
278
append('<inventory%s%s format="5"%s%s>\n' % (
279
fileid1, fileid2, revid1, revid2))
40
e.set('file_id', inv.root.file_id)
41
for path, ie in inv.iter_entries():
42
e.append(self._pack_entry(ie))
46
def _pack_entry(self, ie):
47
"""Convert InventoryEntry to XML element"""
48
assert ie.kind == 'directory' or ie.kind == 'file'
50
e.set('name', ie.name)
51
e.set('file_id', ie.file_id)
53
if ie.text_size != None:
54
e.set('text_size', '%d' % ie.text_size)
56
for f in ['text_version', 'text_sha1', 'entry_version']:
61
# to be conservative, we don't externalize the root pointers
62
# for now, leaving them as null in the xml form. in a future
63
# version it will be implied by nested elements.
64
if ie.parent_id != ROOT_ID:
65
assert isinstance(ie.parent_id, basestring)
66
e.set('parent_id', ie.parent_id)
281
73
def _pack_revision(self, rev):
282
74
"""Revision object -> xml tree"""
283
# For the XML format, we need to write them as Unicode rather than as
284
# utf-8 strings. So that cElementTree can handle properly escaping
286
decode_utf8 = cache_utf8.decode
287
revision_id = rev.revision_id
288
if isinstance(revision_id, str):
289
revision_id = decode_utf8(revision_id)
290
75
root = Element('revision',
291
76
committer = rev.committer,
292
timestamp = '%.3f' % rev.timestamp,
293
revision_id = revision_id,
77
timestamp = '%.9f' % rev.timestamp,
78
revision_id = rev.revision_id,
294
79
inventory_sha1 = rev.inventory_sha1,
297
if rev.timezone is not None:
298
82
root.set('timezone', str(rev.timezone))
300
85
msg = SubElement(root, 'message')
301
86
msg.text = rev.message
304
90
pelts = SubElement(root, 'parents')
305
91
pelts.tail = pelts.text = '\n'
306
for parent_id in rev.parent_ids:
307
assert isinstance(parent_id, basestring)
308
_mod_revision.check_not_reserved_id(parent_id)
92
for rr in rev.parents:
93
assert isinstance(rr, RevisionReference)
309
94
p = SubElement(pelts, 'revision_ref')
311
if isinstance(parent_id, str):
312
parent_id = decode_utf8(parent_id)
313
p.set('revision_id', parent_id)
315
self._pack_revision_properties(rev, root)
97
p.set('revision_id', rr.revision_id)
318
def _pack_revision_properties(self, rev, under_element):
319
top_elt = SubElement(under_element, 'properties')
320
for prop_name, prop_value in sorted(rev.properties.items()):
321
assert isinstance(prop_name, basestring)
322
assert isinstance(prop_value, basestring)
323
prop_elt = SubElement(top_elt, 'property')
324
prop_elt.set('name', prop_name)
325
prop_elt.text = prop_value
329
103
def _unpack_inventory(self, elt):
330
104
"""Construct from XML Element
332
106
assert elt.tag == 'inventory'
333
107
root_id = elt.get('file_id') or ROOT_ID
334
root_id = _get_utf8_or_ascii(root_id)
336
format = elt.get('format')
337
if format is not None:
339
raise BzrError("invalid format version %r on inventory"
341
revision_id = elt.get('revision_id')
342
if revision_id is not None:
343
revision_id = cache_utf8.encode(revision_id)
344
inv = Inventory(root_id, revision_id=revision_id)
108
inv = Inventory(root_id)
346
110
ie = self._unpack_entry(e)
347
if ie.parent_id is None:
111
if ie.parent_id == ROOT_ID:
348
112
ie.parent_id = root_id
352
117
def _unpack_entry(self, elt):
354
if not InventoryEntry.versionable_kind(kind):
355
raise AssertionError('unsupported entry kind %s' % kind)
357
get_cached = _get_utf8_or_ascii
119
assert kind == 'directory' or kind == 'file'
359
121
parent_id = elt.get('parent_id')
360
if parent_id is not None:
361
parent_id = get_cached(parent_id)
362
file_id = get_cached(elt.get('file_id'))
122
if parent_id == None:
364
if kind == 'directory':
365
ie = inventory.InventoryDirectory(file_id,
369
ie = inventory.InventoryFile(file_id,
372
ie.text_sha1 = elt.get('text_sha1')
373
if elt.get('executable') == 'yes':
375
v = elt.get('text_size')
376
ie.text_size = v and int(v)
377
elif kind == 'symlink':
378
ie = inventory.InventoryLink(file_id,
381
ie.symlink_target = elt.get('symlink_target')
383
raise errors.UnsupportedInventoryKind(kind)
384
revision = elt.get('revision')
385
if revision is not None:
386
revision = get_cached(revision)
387
ie.revision = revision
125
ie = InventoryEntry(elt.get('file_id'),
129
ie.text_version = elt.get('text_version')
130
ie.entry_version = elt.get('entry_version')
131
ie.text_sha1 = elt.get('text_sha1')
132
v = elt.get('text_size')
133
ie.text_size = v and int(v)
391
138
def _unpack_revision(self, elt):
392
139
"""XML Element -> Revision object"""
393
140
assert elt.tag == 'revision'
394
format = elt.get('format')
395
if format is not None:
397
raise BzrError("invalid format version %r on inventory"
399
get_cached = _get_utf8_or_ascii
400
142
rev = Revision(committer = elt.get('committer'),
401
143
timestamp = float(elt.get('timestamp')),
402
revision_id = get_cached(elt.get('revision_id')),
144
revision_id = elt.get('revision_id'),
403
145
inventory_sha1 = elt.get('inventory_sha1')
405
148
parents = elt.find('parents') or []
406
149
for p in parents:
407
150
assert p.tag == 'revision_ref', \
408
151
"bad parent node tag %r" % p.tag
409
rev.parent_ids.append(get_cached(p.get('revision_id')))
410
self._unpack_revision_properties(elt, rev)
152
rev_ref = RevisionReference(p.get('revision_id'))
153
rev.parents.append(rev_ref)
411
155
v = elt.get('timezone')
415
rev.timezone = int(v)
156
rev.timezone = v and int(v)
416
158
rev.message = elt.findtext('message') # text of <message>
419
def _unpack_revision_properties(self, elt, rev):
420
"""Unpack properties onto a revision."""
421
props_elt = elt.find('properties')
422
assert len(rev.properties) == 0
425
for prop_elt in props_elt:
426
assert prop_elt.tag == 'property', \
427
"bad tag under properties list: %r" % prop_elt.tag
428
name = prop_elt.get('name')
429
value = prop_elt.text
430
# If a property had an empty value ('') cElementTree reads
431
# that back as None, convert it back to '', so that all
432
# properties have string values
435
assert name not in rev.properties, \
436
"repeated property %r" % name
437
rev.properties[name] = value
440
163
serializer_v5 = Serializer_v5()
added
removed
bzrlib/xml5.py
