← Back to branch summary

~bzr-pqm/bzr/bzr.dev

« back to all changes in this revision

Viewing changes to bzrlib/xml_serializer.py

  • Committer: Naoki INADA
  • Date: 2009-10-29 10:01:19 UTC
  • mto: (4634.97.3 2.0)
  • mto: This revision was merged to the branch mainline in revision 4798.
  • Revision ID: inada-n@klab.jp-20091029100119-uckv9t7ej2qrghw3
import doc-ja rev90

Show diffs side-by-side

added added

removed removed

Lines of Context:
bzrlib/xml_serializer.py
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 import registry
 
25
from bzrlib.serializer import Serializer
26
26
from bzrlib.trace import mutter, warning
27
27
 
28
28
try:
34
34
    except ImportError:
35
35
        from cElementTree import (ElementTree, SubElement, Element,
36
36
                                  XMLTreeBuilder, fromstring, tostring)
37
 
        import elementtree
 
37
        import elementtree.ElementTree
38
38
    ParseError = SyntaxError
39
39
except ImportError:
40
40
    mutter('WARNING: using slower ElementTree; consider installing cElementTree'
49
49
from bzrlib import errors
50
50
 
51
51
 
52
 
class Serializer(object):
53
 
    """Abstract object serialize/deserialize"""
54
 
 
55
 
    def write_inventory(self, inv, f):
56
 
        """Write inventory to a file"""
57
 
        raise NotImplementedError(self.write_inventory)
58
 
 
59
 
    def write_inventory_to_string(self, inv):
60
 
        raise NotImplementedError(self.write_inventory_to_string)
61
 
 
62
 
    def read_inventory_from_string(self, xml_string, revision_id=None):
 
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):
63
59
        """Read xml_string into an inventory object.
64
60
 
65
61
        :param xml_string: The xml to read.
66
62
        :param revision_id: If not-None, the expected revision id of the
67
63
            inventory. Some serialisers use this to set the results' root
68
 
            revision.
 
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.
69
72
        """
70
73
        try:
71
 
            return self._unpack_inventory(fromstring(xml_string), revision_id)
 
74
            return self._unpack_inventory(fromstring(xml_string), revision_id,
 
75
                                          entry_cache=entry_cache)
72
76
        except ParseError, e:
73
77
            raise errors.UnexpectedInventoryFormat(e)
74
78
 
113
117
    }
114
118
def _escape_replace(match, map=escape_map):
115
119
    return map[match.group()]
116
 
 
 
120
 
117
121
def _escape_attrib(text, encoding=None, replace=None):
118
122
    # escape attribute value
119
123
    try:
144
148
    }
145
149
def _escape_cdata_replace(match, map=escape_cdata_map):
146
150
    return map[match.group()]
147
 
 
 
151
 
148
152
def _escape_cdata(text, encoding=None, replace=None):
149
153
    # escape character data
150
154
    try:
166
170
elementtree.ElementTree._escape_cdata = _escape_cdata
167
171
 
168
172
 
169
 
class SerializerRegistry(registry.Registry):
170
 
    """Registry for serializer objects"""
171
 
 
172
 
 
173
 
format_registry = SerializerRegistry()
174
 
format_registry.register_lazy('4', 'bzrlib.xml4', 'serializer_v4')
175
 
format_registry.register_lazy('5', 'bzrlib.xml5', 'serializer_v5')
176
 
format_registry.register_lazy('6', 'bzrlib.xml6', 'serializer_v6')
177
 
format_registry.register_lazy('7', 'bzrlib.xml7', 'serializer_v7')
 
173
def escape_invalid_chars(message):
 
174
    """Escape the XML-invalid characters in a commit message.
 
175
 
 
176
    :param message: Commit message to escape
 
177
    :return: tuple with escaped message and number of characters escaped
 
178
    """
 
179
    if message is None:
 
180
        return None, 0
 
181
    # Python strings can include characters that can't be
 
182
    # represented in well-formed XML; escape characters that
 
183
    # aren't listed in the XML specification
 
184
    # (http://www.w3.org/TR/REC-xml/#NT-Char).
 
185
    return re.subn(u'[^\x09\x0A\x0D\u0020-\uD7FF\uE000-\uFFFD]+',
 
186
            lambda match: match.group(0).encode('unicode_escape'),
 
187
            message)