1
# Copyright (C) 2005, 2006 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
from bzrlib.xml_serializer import SubElement, Element, Serializer
18
from bzrlib.xml import ElementTree, SubElement, Element, Serializer
25
19
from bzrlib.inventory import ROOT_ID, Inventory, InventoryEntry
26
from bzrlib.revision import Revision
20
from bzrlib.revision import Revision
27
21
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
88
"""Clean out the unicode => escaped map"""
89
_unicode_to_escaped_map.clear()
92
27
class Serializer_v5(Serializer):
100
def write_inventory_to_string(self, inv):
101
"""Just call write_inventory with a StringIO and return the value"""
102
sio = cStringIO.StringIO()
103
self.write_inventory(inv, sio)
104
return sio.getvalue()
106
def write_inventory(self, inv, f):
107
"""Write inventory to a file.
109
:param inv: the inventory to write.
110
:param f: the file to write.
114
append = output.append
115
self._append_inventory_root(append, inv)
116
entries = inv.iter_entries()
118
root_path, root_ie = entries.next()
119
for path, ie in entries:
120
self._append_entry(append, ie)
121
append('</inventory>\n')
123
# Just to keep the cache from growing without bounds
124
# but we may actually not want to do clear the cache
127
def _append_inventory_root(self, append, inv):
128
"""Append the inventory root to output."""
35
def _pack_inventory(self, inv):
36
"""Convert to XML Element"""
37
e = Element('inventory')
130
39
if inv.root.file_id not in (None, ROOT_ID):
132
append(_encode_and_escape(inv.root.file_id))
133
append(' format="5"')
134
if inv.revision_id is not None:
135
append(' revision_id="')
136
append(_encode_and_escape(inv.revision_id))
139
def _append_entry(self, append, ie):
140
"""Convert InventoryEntry to XML element and append to output."""
141
# TODO: should just be a plain assertion
142
assert InventoryEntry.versionable_kind(ie.kind), \
143
'unsupported entry kind %s' % ie.kind
148
append(' executable="yes"')
150
append(_encode_and_escape(ie.file_id))
152
append(_encode_and_escape(ie.name))
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 in ('directory', 'file', 'symlink')
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_sha1', 'revision', 'symlink_target']:
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.
153
64
if ie.parent_id != ROOT_ID:
154
65
assert isinstance(ie.parent_id, basestring)
155
append(' parent_id="')
156
append(_encode_and_escape(ie.parent_id))
157
if ie.revision is not None:
158
append(' revision="')
159
append(_encode_and_escape(ie.revision))
160
if ie.symlink_target is not None:
161
append(' symlink_target="')
162
append(_encode_and_escape(ie.symlink_target))
163
if ie.text_sha1 is not None:
164
append(' text_sha1="')
167
if ie.text_size is not None:
168
append(' text_size="%d"' % ie.text_size)
66
e.set('parent_id', ie.parent_id)
172
73
def _pack_revision(self, rev):
173
74
"""Revision object -> xml tree"""
192
94
p = SubElement(pelts, 'revision_ref')
194
96
p.set('revision_id', parent_id)
196
self._pack_revision_properties(rev, root)
199
def _pack_revision_properties(self, rev, under_element):
200
top_elt = SubElement(under_element, 'properties')
201
for prop_name, prop_value in sorted(rev.properties.items()):
202
assert isinstance(prop_name, basestring)
203
assert isinstance(prop_value, basestring)
204
prop_elt = SubElement(top_elt, 'property')
205
prop_elt.set('name', prop_name)
206
prop_elt.text = prop_value
210
101
def _unpack_inventory(self, elt):
211
102
"""Construct from XML Element
213
104
assert elt.tag == 'inventory'
214
105
root_id = elt.get('file_id') or ROOT_ID
215
format = elt.get('format')
216
if format is not None:
218
raise BzrError("invalid format version %r on inventory"
220
revision_id = elt.get('revision_id')
221
if revision_id is not None:
222
revision_id = cache_utf8.get_cached_unicode(revision_id)
223
inv = Inventory(root_id, revision_id=revision_id)
106
inv = Inventory(root_id)
225
108
ie = self._unpack_entry(e)
226
109
if ie.parent_id == ROOT_ID:
231
115
def _unpack_entry(self, elt):
233
if not InventoryEntry.versionable_kind(kind):
117
if not kind in ('directory', 'file', 'symlink'):
234
118
raise AssertionError('unsupported entry kind %s' % kind)
236
get_cached = cache_utf8.get_cached_unicode
238
120
parent_id = elt.get('parent_id')
239
121
if parent_id == None:
240
122
parent_id = ROOT_ID
241
# TODO: jam 20060817 At present, caching file ids costs us too
242
# much time. It slows down overall read performances from
243
# approx 500ms to 700ms. And doesn't improve future reads.
244
# it might be because revision ids and file ids are mixing.
245
# Consider caching *just* the file ids, for a limited period
247
#parent_id = get_cached(parent_id)
248
#file_id = get_cached(elt.get('file_id'))
249
file_id = elt.get('file_id')
251
if kind == 'directory':
252
ie = inventory.InventoryDirectory(file_id,
256
ie = inventory.InventoryFile(file_id,
259
ie.text_sha1 = elt.get('text_sha1')
260
if elt.get('executable') == 'yes':
262
v = elt.get('text_size')
263
ie.text_size = v and int(v)
264
elif kind == 'symlink':
265
ie = inventory.InventoryLink(file_id,
268
ie.symlink_target = elt.get('symlink_target')
270
raise BzrError("unknown kind %r" % kind)
271
revision = elt.get('revision')
272
if revision is not None:
273
revision = get_cached(revision)
274
ie.revision = revision
124
ie = InventoryEntry(elt.get('file_id'),
128
ie.revision = elt.get('revision')
129
ie.text_sha1 = elt.get('text_sha1')
130
ie.symlink_target = elt.get('symlink_target')
131
v = elt.get('text_size')
132
ie.text_size = v and int(v)
278
137
def _unpack_revision(self, elt):
279
138
"""XML Element -> Revision object"""
280
139
assert elt.tag == 'revision'
281
format = elt.get('format')
282
if format is not None:
284
raise BzrError("invalid format version %r on inventory"
286
get_cached = cache_utf8.get_cached_unicode
287
141
rev = Revision(committer = elt.get('committer'),
288
142
timestamp = float(elt.get('timestamp')),
289
revision_id = get_cached(elt.get('revision_id')),
143
revision_id = elt.get('revision_id'),
290
144
inventory_sha1 = elt.get('inventory_sha1')
292
147
parents = elt.find('parents') or []
293
148
for p in parents:
294
149
assert p.tag == 'revision_ref', \
295
150
"bad parent node tag %r" % p.tag
296
rev.parent_ids.append(get_cached(p.get('revision_id')))
297
self._unpack_revision_properties(elt, rev)
151
rev.parent_ids.append(p.get('revision_id'))
298
153
v = elt.get('timezone')
302
rev.timezone = int(v)
154
rev.timezone = v and int(v)
303
156
rev.message = elt.findtext('message') # text of <message>
306
def _unpack_revision_properties(self, elt, rev):
307
"""Unpack properties onto a revision."""
308
props_elt = elt.find('properties')
309
assert len(rev.properties) == 0
312
for prop_elt in props_elt:
313
assert prop_elt.tag == 'property', \
314
"bad tag under properties list: %r" % prop_elt.tag
315
name = prop_elt.get('name')
316
value = prop_elt.text
317
# If a property had an empty value ('') cElementTree reads
318
# that back as None, convert it back to '', so that all
319
# properties have string values
322
assert name not in rev.properties, \
323
"repeated property %r" % name
324
rev.properties[name] = value
327
161
serializer_v5 = Serializer_v5()
added
removed
bzrlib/xml5.py
