← Back to branch summary

~bzr-pqm/bzr/bzr.dev

« back to all changes in this revision

Viewing changes to bzrlib/xml_serializer.py

  • Committer: Marius Kruger
  • Date: 2010-07-10 21:28:56 UTC
  • mto: (5384.1.1 integration)
  • mto: This revision was merged to the branch mainline in revision 5385.
  • Revision ID: marius.kruger@enerweb.co.za-20100710212856-uq4ji3go0u5se7hx
* Update documentation
* add NEWS

Show diffs side-by-side

added added

removed removed

Lines of Context:
bzrlib/xml_serializer.py
1
 
# Copyright (C) 2005, 2006 Canonical Ltd
 
1
# Copyright (C) 2005-2010 Canonical Ltd
2
2
#
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
12
12
#
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
 
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
 
15
# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
16
16
 
17
17
"""XML externalization support."""
18
18
 
22
22
# importing this module is fairly slow because it has to load several
23
23
# ElementTree bits
24
24
 
25
 
from bzrlib.trace import mutter, warning
 
25
from bzrlib.serializer import Serializer
 
26
from bzrlib.trace import mutter
26
27
 
27
28
try:
28
29
    try:
33
34
    except ImportError:
34
35
        from cElementTree import (ElementTree, SubElement, Element,
35
36
                                  XMLTreeBuilder, fromstring, tostring)
36
 
        import elementtree
 
37
        import elementtree.ElementTree
37
38
    ParseError = SyntaxError
38
39
except ImportError:
39
40
    mutter('WARNING: using slower ElementTree; consider installing cElementTree'
48
49
from bzrlib import errors
49
50
 
50
51
 
51
 
class Serializer(object):
52
 
    """Abstract object serialize/deserialize"""
53
 
    def write_inventory(self, inv, f):
54
 
        """Write inventory to a file"""
55
 
        elt = self._pack_inventory(inv)
56
 
        self._write_element(elt, f)
57
 
 
58
 
    def write_inventory_to_string(self, inv):
59
 
        return tostring(self._pack_inventory(inv)) + '\n'
60
 
 
61
 
    def read_inventory_from_string(self, xml_string):
 
52
class XMLSerializer(Serializer):
 
53
    """Abstract XML object serialize/deserialize"""
 
54
 
 
55
    squashes_xml_invalid_characters = True
 
56
 
 
57
    def read_inventory_from_string(self, xml_string, revision_id=None,
 
58
                                   entry_cache=None, return_from_cache=False):
 
59
        """Read xml_string into an inventory object.
 
60
 
 
61
        :param xml_string: The xml to read.
 
62
        :param revision_id: If not-None, the expected revision id of the
 
63
            inventory. Some serialisers use this to set the results' root
 
64
            revision. This should be supplied for deserialising all
 
65
            from-repository inventories so that xml5 inventories that were
 
66
            serialised without a revision identifier can be given the right
 
67
            revision id (but not for working tree inventories where users can
 
68
            edit the data without triggering checksum errors or anything).
 
69
        :param entry_cache: An optional cache of InventoryEntry objects. If
 
70
            supplied we will look up entries via (file_id, revision_id) which
 
71
            should map to a valid InventoryEntry (File/Directory/etc) object.
 
72
        :param return_from_cache: Return entries directly from the cache,
 
73
            rather than copying them first. This is only safe if the caller
 
74
            promises not to mutate the returned inventory entries, but it can
 
75
            make some operations significantly faster.
 
76
        """
62
77
        try:
63
 
            return self._unpack_inventory(fromstring(xml_string))
 
78
            return self._unpack_inventory(fromstring(xml_string), revision_id,
 
79
                                          entry_cache=entry_cache,
 
80
                                          return_from_cache=return_from_cache)
64
81
        except ParseError, e:
65
82
            raise errors.UnexpectedInventoryFormat(e)
66
83
 
67
 
    def read_inventory(self, f):
 
84
    def read_inventory(self, f, revision_id=None):
68
85
        try:
69
 
            return self._unpack_inventory(self._read_element(f))
 
86
            try:
 
87
                return self._unpack_inventory(self._read_element(f),
 
88
                    revision_id=None)
 
89
            finally:
 
90
                f.close()
70
91
        except ParseError, e:
71
92
            raise errors.UnexpectedInventoryFormat(e)
72
93
 
104
125
    }
105
126
def _escape_replace(match, map=escape_map):
106
127
    return map[match.group()]
107
 
 
 
128
 
108
129
def _escape_attrib(text, encoding=None, replace=None):
109
130
    # escape attribute value
110
131
    try:
135
156
    }
136
157
def _escape_cdata_replace(match, map=escape_cdata_map):
137
158
    return map[match.group()]
138
 
 
 
159
 
139
160
def _escape_cdata(text, encoding=None, replace=None):
140
161
    # escape character data
141
162
    try:
155
176
        elementtree.ElementTree._raise_serialization_error(text)
156
177
 
157
178
elementtree.ElementTree._escape_cdata = _escape_cdata
 
179
 
 
180
 
 
181
def escape_invalid_chars(message):
 
182
    """Escape the XML-invalid characters in a commit message.
 
183
 
 
184
    :param message: Commit message to escape
 
185
    :return: tuple with escaped message and number of characters escaped
 
186
    """
 
187
    if message is None:
 
188
        return None, 0
 
189
    # Python strings can include characters that can't be
 
190
    # represented in well-formed XML; escape characters that
 
191
    # aren't listed in the XML specification
 
192
    # (http://www.w3.org/TR/REC-xml/#NT-Char).
 
193
    return re.subn(u'[^\x09\x0A\x0D\u0020-\uD7FF\uE000-\uFFFD]+',
 
194
            lambda match: match.group(0).encode('unicode_escape'),
 
195
            message)