~bzr-pqm/bzr/bzr.dev

1 by mbp at sourcefrog
import from baz patch-364
1
#! /usr/bin/env python
2
# -*- coding: UTF-8 -*-
3
4
# This program is free software; you can redistribute it and/or modify
5
# it under the terms of the GNU General Public License as published by
6
# the Free Software Foundation; either version 2 of the License, or
7
# (at your option) any later version.
8
9
# This program is distributed in the hope that it will be useful,
10
# but WITHOUT ANY WARRANTY; without even the implied warranty of
11
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
12
# GNU General Public License for more details.
13
14
# You should have received a copy of the GNU General Public License
15
# along with this program; if not, write to the Free Software
16
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
17
18
"""XML externalization support."""
19
48 by Martin Pool
witty comment
20
# "XML is like violence: if it doesn't solve your problem, you aren't
21
# using enough of it." -- various
22
1180 by Martin Pool
- start splitting code for xml (de)serialization away from objects
23
# importing this module is fairly slow because it has to load several
24
# ElementTree bits
25
802 by Martin Pool
- Remove XMLMixin class in favour of simple pack_xml, unpack_xml functions
26
try:
974.1.26 by aaron.bentley at utoronto
merged mbp@sourcefrog.net-20050817233101-0939da1cf91f2472
27
    from util.cElementTree import ElementTree, SubElement, Element
802 by Martin Pool
- Remove XMLMixin class in favour of simple pack_xml, unpack_xml functions
28
except ImportError:
974.1.26 by aaron.bentley at utoronto
merged mbp@sourcefrog.net-20050817233101-0939da1cf91f2472
29
    from util.elementtree.ElementTree import ElementTree, SubElement, Element
802 by Martin Pool
- Remove XMLMixin class in favour of simple pack_xml, unpack_xml functions
30
1180 by Martin Pool
- start splitting code for xml (de)serialization away from objects
31
from bzrlib.inventory import ROOT_ID, Inventory, InventoryEntry
1182 by Martin Pool
- more disentangling of xml storage format from objects
32
from bzrlib.revision import Revision, RevisionReference        
1183 by Martin Pool
- implement version 5 xml storage, and tests
33
from bzrlib.errors import BzrError
1180 by Martin Pool
- start splitting code for xml (de)serialization away from objects
34
35
36
class Serializer(object):
37
    """Abstract object serialize/deserialize"""
38
    def write_inventory(self, inv, f):
39
        """Write inventory to a file"""
40
        elt = self._pack_inventory(inv)
41
        self._write_element(elt, f)
42
43
    def read_inventory(self, f):
44
        return self._unpack_inventory(self._read_element(f))
45
1182 by Martin Pool
- more disentangling of xml storage format from objects
46
    def write_revision(self, rev, f):
47
        self._write_element(self._pack_revision(rev), f)
48
49
    def read_revision(self, f):
50
        return self._unpack_revision(self._read_element(f))
51
1180 by Martin Pool
- start splitting code for xml (de)serialization away from objects
52
    def _write_element(self, elt, f):
53
        ElementTree(elt).write(f, 'utf-8')
54
        f.write('\n')
55
56
    def _read_element(self, f):
57
        return ElementTree().parse(f)
58
59
60
61
class _Serializer_v4(Serializer):
1183 by Martin Pool
- implement version 5 xml storage, and tests
62
    """Version 0.0.4 serializer
63
64
    You should use the serialzer_v4 singleton."""
1180 by Martin Pool
- start splitting code for xml (de)serialization away from objects
65
    
66
    __slots__ = []
67
    
68
    def _pack_inventory(self, inv):
69
        """Convert to XML Element"""
70
        e = Element('inventory')
71
        e.text = '\n'
72
        if inv.root.file_id not in (None, ROOT_ID):
73
            e.set('file_id', inv.root.file_id)
74
        for path, ie in inv.iter_entries():
75
            e.append(self._pack_entry(ie))
76
        return e
77
78
79
    def _pack_entry(self, ie):
80
        """Convert InventoryEntry to XML element"""
81
        e = Element('entry')
82
        e.set('name', ie.name)
83
        e.set('file_id', ie.file_id)
84
        e.set('kind', ie.kind)
85
86
        if ie.text_size != None:
87
            e.set('text_size', '%d' % ie.text_size)
88
89
        for f in ['text_id', 'text_sha1']:
90
            v = getattr(ie, f)
91
            if v != None:
92
                e.set(f, v)
93
94
        # to be conservative, we don't externalize the root pointers
95
        # for now, leaving them as null in the xml form.  in a future
96
        # version it will be implied by nested elements.
97
        if ie.parent_id != ROOT_ID:
98
            assert isinstance(ie.parent_id, basestring)
99
            e.set('parent_id', ie.parent_id)
100
101
        e.tail = '\n'
102
103
        return e
104
105
106
    def _unpack_inventory(self, elt):
107
        """Construct from XML Element
108
        """
109
        assert elt.tag == 'inventory'
110
        root_id = elt.get('file_id') or ROOT_ID
111
        inv = Inventory(root_id)
112
        for e in elt:
113
            ie = self._unpack_entry(e)
114
            if ie.parent_id == ROOT_ID:
115
                ie.parent_id = root_id
116
            inv.add(ie)
117
        return inv
118
119
120
    def _unpack_entry(self, elt):
121
        assert elt.tag == 'entry'
122
123
        ## original format inventories don't have a parent_id for
124
        ## nodes in the root directory, but it's cleaner to use one
125
        ## internally.
126
        parent_id = elt.get('parent_id')
127
        if parent_id == None:
128
            parent_id = ROOT_ID
129
130
        ie = InventoryEntry(elt.get('file_id'),
131
                              elt.get('name'),
132
                              elt.get('kind'),
133
                              parent_id)
134
        ie.text_id = elt.get('text_id')
135
        ie.text_sha1 = elt.get('text_sha1')
136
137
        ## mutter("read inventoryentry: %r" % (elt.attrib))
138
139
        v = elt.get('text_size')
140
        ie.text_size = v and int(v)
141
142
        return ie
143
144
1182 by Martin Pool
- more disentangling of xml storage format from objects
145
    def _pack_revision(self, rev):
146
        """Revision object -> xml tree"""
147
        root = Element('revision',
148
                       committer = rev.committer,
149
                       timestamp = '%.9f' % rev.timestamp,
150
                       revision_id = rev.revision_id,
151
                       inventory_id = rev.inventory_id,
152
                       inventory_sha1 = rev.inventory_sha1,
153
                       )
154
        if rev.timezone:
155
            root.set('timezone', str(rev.timezone))
156
        root.text = '\n'
157
158
        msg = SubElement(root, 'message')
159
        msg.text = rev.message
160
        msg.tail = '\n'
161
162
        if rev.parents:
163
            pelts = SubElement(root, 'parents')
164
            pelts.tail = pelts.text = '\n'
165
            for rr in rev.parents:
166
                assert isinstance(rr, RevisionReference)
167
                p = SubElement(pelts, 'revision_ref')
168
                p.tail = '\n'
169
                assert rr.revision_id
170
                p.set('revision_id', rr.revision_id)
171
                if rr.revision_sha1:
172
                    p.set('revision_sha1', rr.revision_sha1)
173
174
        return root
175
176
    
177
    def _unpack_revision(self, elt):
178
        """XML Element -> Revision object"""
179
        
180
        # <changeset> is deprecated...
181
        if elt.tag not in ('revision', 'changeset'):
1183 by Martin Pool
- implement version 5 xml storage, and tests
182
            raise BzrError("unexpected tag in revision file: %r" % elt)
1182 by Martin Pool
- more disentangling of xml storage format from objects
183
184
        rev = Revision(committer = elt.get('committer'),
185
                       timestamp = float(elt.get('timestamp')),
186
                       revision_id = elt.get('revision_id'),
187
                       inventory_id = elt.get('inventory_id'),
188
                       inventory_sha1 = elt.get('inventory_sha1')
189
                       )
190
191
        precursor = elt.get('precursor')
192
        precursor_sha1 = elt.get('precursor_sha1')
193
194
        pelts = elt.find('parents')
195
196
        if pelts:
197
            for p in pelts:
198
                assert p.tag == 'revision_ref', \
199
                       "bad parent node tag %r" % p.tag
200
                rev_ref = RevisionReference(p.get('revision_id'),
201
                                            p.get('revision_sha1'))
202
                rev.parents.append(rev_ref)
203
204
            if precursor:
205
                # must be consistent
206
                prec_parent = rev.parents[0].revision_id
207
                assert prec_parent == precursor
208
        elif precursor:
209
            # revisions written prior to 0.0.5 have a single precursor
210
            # give as an attribute
211
            rev_ref = RevisionReference(precursor, precursor_sha1)
212
            rev.parents.append(rev_ref)
213
214
        v = elt.get('timezone')
215
        rev.timezone = v and int(v)
216
217
        rev.message = elt.findtext('message') # text of <message>
218
        return rev
219
220
221
1183 by Martin Pool
- implement version 5 xml storage, and tests
222
class _Serializer_v5(Serializer):
223
    """Version 5 serializer
224
1184 by Martin Pool
- fix v5 packing of inventory entries
225
    Packs objects into XML and vice versa.
226
1183 by Martin Pool
- implement version 5 xml storage, and tests
227
    You should use the serialzer_v5 singleton."""
228
    
229
    __slots__ = []
230
    
231
    def _pack_inventory(self, inv):
232
        """Convert to XML Element"""
233
        e = Element('inventory')
234
        e.text = '\n'
235
        if inv.root.file_id not in (None, ROOT_ID):
236
            e.set('file_id', inv.root.file_id)
237
        for path, ie in inv.iter_entries():
238
            e.append(self._pack_entry(ie))
239
        return e
240
241
242
    def _pack_entry(self, ie):
243
        """Convert InventoryEntry to XML element"""
1184 by Martin Pool
- fix v5 packing of inventory entries
244
        assert ie.kind == 'directory' or ie.kind == 'file'
245
        e = Element(ie.kind)
1183 by Martin Pool
- implement version 5 xml storage, and tests
246
        e.set('name', ie.name)
247
        e.set('file_id', ie.file_id)
248
249
        if ie.text_size != None:
250
            e.set('text_size', '%d' % ie.text_size)
251
252
        for f in ['text_version', 'text_sha1', 'entry_version']:
253
            v = getattr(ie, f)
254
            if v != None:
255
                e.set(f, v)
256
257
        # to be conservative, we don't externalize the root pointers
258
        # for now, leaving them as null in the xml form.  in a future
259
        # version it will be implied by nested elements.
260
        if ie.parent_id != ROOT_ID:
261
            assert isinstance(ie.parent_id, basestring)
262
            e.set('parent_id', ie.parent_id)
263
264
        e.tail = '\n'
265
266
        return e
267
268
269
    def _pack_revision(self, rev):
270
        """Revision object -> xml tree"""
271
        root = Element('revision',
272
                       committer = rev.committer,
273
                       timestamp = '%.9f' % rev.timestamp,
274
                       revision_id = rev.revision_id,
275
                       inventory_id = rev.inventory_id,
276
                       inventory_sha1 = rev.inventory_sha1,
277
                       )
278
        if rev.timezone:
279
            root.set('timezone', str(rev.timezone))
280
        root.text = '\n'
281
282
        msg = SubElement(root, 'message')
283
        msg.text = rev.message
284
        msg.tail = '\n'
285
286
        if rev.parents:
287
            pelts = SubElement(root, 'parents')
288
            pelts.tail = pelts.text = '\n'
289
            for rr in rev.parents:
290
                assert isinstance(rr, RevisionReference)
291
                p = SubElement(pelts, 'revision_ref')
292
                p.tail = '\n'
293
                assert rr.revision_id
294
                p.set('revision_id', rr.revision_id)
295
296
        return root
297
298
    
299
300
    def _unpack_inventory(self, elt):
301
        """Construct from XML Element
302
        """
303
        assert elt.tag == 'inventory'
304
        root_id = elt.get('file_id') or ROOT_ID
305
        inv = Inventory(root_id)
306
        for e in elt:
307
            ie = self._unpack_entry(e)
308
            if ie.parent_id == ROOT_ID:
309
                ie.parent_id = root_id
310
            inv.add(ie)
311
        return inv
312
313
314
    def _unpack_entry(self, elt):
315
        kind = elt.tag
316
        assert kind == 'directory' or kind == 'file'
317
318
        parent_id = elt.get('parent_id')
319
        if parent_id == None:
320
            parent_id = ROOT_ID
321
322
        ie = InventoryEntry(elt.get('file_id'),
323
                            elt.get('name'),
324
                            kind,
325
                            parent_id)
326
        ie.text_version = elt.get('text_version')
327
        ie.entry_version = elt.get('entry_version')
328
        ie.text_sha1 = elt.get('text_sha1')
329
        v = elt.get('text_size')
330
        ie.text_size = v and int(v)
331
332
        return ie
333
334
335
    def _unpack_revision(self, elt):
336
        """XML Element -> Revision object"""
337
        assert elt.tag == 'revision'
338
        
339
        rev = Revision(committer = elt.get('committer'),
340
                       timestamp = float(elt.get('timestamp')),
341
                       revision_id = elt.get('revision_id'),
342
                       inventory_id = elt.get('inventory_id'),
343
                       inventory_sha1 = elt.get('inventory_sha1')
344
                       )
345
346
        for p in elt.find('parents'):
347
            assert p.tag == 'revision_ref', \
348
                   "bad parent node tag %r" % p.tag
349
            rev_ref = RevisionReference(p.get('revision_id'))
350
            rev.parents.append(rev_ref)
351
352
        v = elt.get('timezone')
353
        rev.timezone = v and int(v)
354
355
        rev.message = elt.findtext('message') # text of <message>
356
        return rev
357
358
359
1180 by Martin Pool
- start splitting code for xml (de)serialization away from objects
360
"""singleton instance"""
361
serializer_v4 = _Serializer_v4()
1183 by Martin Pool
- implement version 5 xml storage, and tests
362
363
serializer_v5 = _Serializer_v5()