aboutsummaryrefslogtreecommitdiff
path: root/tests/python_dependencies/impacket/structure.py
blob: 7a04117ff8ba722ded0867e93212740c9dbafe94 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
# Copyright (c) 2003-2016 CORE Security Technologies
#
# This software is provided under under a slightly modified version
# of the Apache Software License. See the accompanying LICENSE file
# for more information.
#

from struct import pack, unpack, calcsize

class Structure:
    """ sublcasses can define commonHdr and/or structure.
        each of them is an tuple of either two: (fieldName, format) or three: (fieldName, ':', class) fields.
        [it can't be a dictionary, because order is important]
        
        where format specifies how the data in the field will be converted to/from bytes (string)
        class is the class to use when unpacking ':' fields.

        each field can only contain one value (or an array of values for *)
           i.e. struct.pack('Hl',1,2) is valid, but format specifier 'Hl' is not (you must use 2 dfferent fields)

        format specifiers:
          specifiers from module pack can be used with the same format 
          see struct.__doc__ (pack/unpack is finally called)
            x       [padding byte]
            c       [character]
            b       [signed byte]
            B       [unsigned byte]
            h       [signed short]
            H       [unsigned short]
            l       [signed long]
            L       [unsigned long]
            i       [signed integer]
            I       [unsigned integer]
            q       [signed long long (quad)]
            Q       [unsigned long long (quad)]
            s       [string (array of chars), must be preceded with length in format specifier, padded with zeros]
            p       [pascal string (includes byte count), must be preceded with length in format specifier, padded with zeros]
            f       [float]
            d       [double]
            =       [native byte ordering, size and alignment]
            @       [native byte ordering, standard size and alignment]
            !       [network byte ordering]
            <       [little endian]
            >       [big endian]

          usual printf like specifiers can be used (if started with %) 
          [not recommeneded, there is no why to unpack this]

            %08x    will output an 8 bytes hex
            %s      will output a string
            %s\\x00  will output a NUL terminated string
            %d%d    will output 2 decimal digits (against the very same specification of Structure)
            ...

          some additional format specifiers:
            :       just copy the bytes from the field into the output string (input may be string, other structure, or anything responding to __str__()) (for unpacking, all what's left is returned)
            z       same as :, but adds a NUL byte at the end (asciiz) (for unpacking the first NUL byte is used as terminator)  [asciiz string]
            u       same as z, but adds two NUL bytes at the end (after padding to an even size with NULs). (same for unpacking) [unicode string]
            w       DCE-RPC/NDR string (it's a macro for [  '<L=(len(field)+1)/2','"\\x00\\x00\\x00\\x00','<L=(len(field)+1)/2',':' ]
            ?-field length of field named 'field', formated as specified with ? ('?' may be '!H' for example). The input value overrides the real length
            ?1*?2   array of elements. Each formated as '?2', the number of elements in the array is stored as specified by '?1' (?1 is optional, or can also be a constant (number), for unpacking)
            'xxxx   literal xxxx (field's value doesn't change the output. quotes must not be closed or escaped)
            "xxxx   literal xxxx (field's value doesn't change the output. quotes must not be closed or escaped)
            _       will not pack the field. Accepts a third argument, which is an unpack code. See _Test_UnpackCode for an example
            ?=packcode  will evaluate packcode in the context of the structure, and pack the result as specified by ?. Unpacking is made plain
            ?&fieldname "Address of field fieldname".
                        For packing it will simply pack the id() of fieldname. Or use 0 if fieldname doesn't exists.
                        For unpacking, it's used to know weather fieldname has to be unpacked or not, i.e. by adding a & field you turn another field (fieldname) in an optional field.
            
    """
    commonHdr = ()
    structure = ()
    debug = 0

    def __init__(self, data = None, alignment = 0):
        if not hasattr(self, 'alignment'):
            self.alignment = alignment

        self.fields    = {}
        self.rawData   = data
        if data is not None:
            self.fromString(data)
        else:
            self.data = None

    @classmethod
    def fromFile(self, file):
        answer = self()
        answer.fromString(file.read(len(answer)))
        return answer

    def setAlignment(self, alignment):
        self.alignment = alignment

    def setData(self, data):
        self.data = data

    def packField(self, fieldName, format = None):
        if self.debug:
            print "packField( %s | %s )" % (fieldName, format)

        if format is None:
            format = self.formatForField(fieldName)

        if self.fields.has_key(fieldName):
            ans = self.pack(format, self.fields[fieldName], field = fieldName)
        else:
            ans = self.pack(format, None, field = fieldName)

        if self.debug:
            print "\tanswer %r" % ans

        return ans

    def getData(self):
        if self.data is not None:
            return self.data
        data = ''
        for field in self.commonHdr+self.structure:
            try:
                data += self.packField(field[0], field[1])
            except Exception, e:
                if self.fields.has_key(field[0]):
                    e.args += ("When packing field '%s | %s | %r' in %s" % (field[0], field[1], self[field[0]], self.__class__),)
                else:
                    e.args += ("When packing field '%s | %s' in %s" % (field[0], field[1], self.__class__),)
                raise
            if self.alignment:
                if len(data) % self.alignment:
                    data += ('\x00'*self.alignment)[:-(len(data) % self.alignment)]
            
        #if len(data) % self.alignment: data += ('\x00'*self.alignment)[:-(len(data) % self.alignment)]
        return data

    def fromString(self, data):
        self.rawData = data
        for field in self.commonHdr+self.structure:
            if self.debug:
                print "fromString( %s | %s | %r )" % (field[0], field[1], data)
            size = self.calcUnpackSize(field[1], data, field[0])
            if self.debug:
                print "  size = %d" % size
            dataClassOrCode = str
            if len(field) > 2:
                dataClassOrCode = field[2]
            try:
                self[field[0]] = self.unpack(field[1], data[:size], dataClassOrCode = dataClassOrCode, field = field[0])
            except Exception,e:
                e.args += ("When unpacking field '%s | %s | %r[:%d]'" % (field[0], field[1], data, size),)
                raise

            size = self.calcPackSize(field[1], self[field[0]], field[0])
            if self.alignment and size % self.alignment:
                size += self.alignment - (size % self.alignment)
            data = data[size:]

        return self
        
    def __setitem__(self, key, value):
        self.fields[key] = value
        self.data = None        # force recompute

    def __getitem__(self, key):
        return self.fields[key]

    def __delitem__(self, key):
        del self.fields[key]
        
    def __str__(self):
        return self.getData()

    def __len__(self):
        # XXX: improve
        return len(self.getData())

    def pack(self, format, data, field = None):
        if self.debug:
            print "  pack( %s | %r | %s)" %  (format, data, field)

        if field:
            addressField = self.findAddressFieldFor(field)
            if (addressField is not None) and (data is None):
                return ''

        # void specifier
        if format[:1] == '_':
            return ''

        # quote specifier
        if format[:1] == "'" or format[:1] == '"':
            return format[1:]

        # code specifier
        two = format.split('=')
        if len(two) >= 2:
            try:
                return self.pack(two[0], data)
            except:
                fields = {'self':self}
                fields.update(self.fields)
                return self.pack(two[0], eval(two[1], {}, fields))

        # address specifier
        two = format.split('&')
        if len(two) == 2:
            try:
                return self.pack(two[0], data)
            except:
                if (self.fields.has_key(two[1])) and (self[two[1]] is not None):
                    return self.pack(two[0], id(self[two[1]]) & ((1<<(calcsize(two[0])*8))-1) )
                else:
                    return self.pack(two[0], 0)

        # length specifier
        two = format.split('-')
        if len(two) == 2:
            try:
                return self.pack(two[0],data)
            except:
                return self.pack(two[0], self.calcPackFieldSize(two[1]))

        # array specifier
        two = format.split('*')
        if len(two) == 2:
            answer = ''
            for each in data:
                answer += self.pack(two[1], each)
            if two[0]:
                if two[0].isdigit():
                    if int(two[0]) != len(data):
                        raise Exception, "Array field has a constant size, and it doesn't match the actual value"
                else:
                    return self.pack(two[0], len(data))+answer
            return answer

        # "printf" string specifier
        if format[:1] == '%':
            # format string like specifier
            return format % data

        # asciiz specifier
        if format[:1] == 'z':
            return str(data)+'\0'

        # unicode specifier
        if format[:1] == 'u':
            return str(data)+'\0\0' + (len(data) & 1 and '\0' or '')

        # DCE-RPC/NDR string specifier
        if format[:1] == 'w':
            if len(data) == 0:
                data = '\0\0'
            elif len(data) % 2:
                data += '\0'
            l = pack('<L', len(data)/2)
            return '%s\0\0\0\0%s%s' % (l,l,data)
                    
        if data is None:
            raise Exception, "Trying to pack None"
        
        # literal specifier
        if format[:1] == ':':
            return str(data)

        # struct like specifier
        return pack(format, data)

    def unpack(self, format, data, dataClassOrCode = str, field = None):
        if self.debug:
            print "  unpack( %s | %r )" %  (format, data)

        if field:
            addressField = self.findAddressFieldFor(field)
            if addressField is not None:
                if not self[addressField]:
                    return

        # void specifier
        if format[:1] == '_':
            if dataClassOrCode != str:
                fields = {'self':self, 'inputDataLeft':data}
                fields.update(self.fields)
                return eval(dataClassOrCode, {}, fields)
            else:
                return None

        # quote specifier
        if format[:1] == "'" or format[:1] == '"':
            answer = format[1:]
            if answer != data:
                raise Exception, "Unpacked data doesn't match constant value '%r' should be '%r'" % (data, answer)
            return answer

        # address specifier
        two = format.split('&')
        if len(two) == 2:
            return self.unpack(two[0],data)

        # code specifier
        two = format.split('=')
        if len(two) >= 2:
            return self.unpack(two[0],data)

        # length specifier
        two = format.split('-')
        if len(two) == 2:
            return self.unpack(two[0],data)

        # array specifier
        two = format.split('*')
        if len(two) == 2:
            answer = []
            sofar = 0
            if two[0].isdigit():
                number = int(two[0])
            elif two[0]:
                sofar += self.calcUnpackSize(two[0], data)
                number = self.unpack(two[0], data[:sofar])
            else:
                number = -1

            while number and sofar < len(data):
                nsofar = sofar + self.calcUnpackSize(two[1],data[sofar:])
                answer.append(self.unpack(two[1], data[sofar:nsofar], dataClassOrCode))
                number -= 1
                sofar = nsofar
            return answer

        # "printf" string specifier
        if format[:1] == '%':
            # format string like specifier
            return format % data

        # asciiz specifier
        if format == 'z':
            if data[-1] != '\x00':
                raise Exception, ("%s 'z' field is not NUL terminated: %r" % (field, data))
            return data[:-1] # remove trailing NUL

        # unicode specifier
        if format == 'u':
            if data[-2:] != '\x00\x00':
                raise Exception, ("%s 'u' field is not NUL-NUL terminated: %r" % (field, data))
            return data[:-2] # remove trailing NUL

        # DCE-RPC/NDR string specifier
        if format == 'w':
            l = unpack('<L', data[:4])[0]
            return data[12:12+l*2]

        # literal specifier
        if format == ':':
            return dataClassOrCode(data)

        # struct like specifier
        return unpack(format, data)[0]

    def calcPackSize(self, format, data, field = None):
#        # print "  calcPackSize  %s:%r" %  (format, data)
        if field:
            addressField = self.findAddressFieldFor(field)
            if addressField is not None:
                if not self[addressField]:
                    return 0

        # void specifier
        if format[:1] == '_':
            return 0

        # quote specifier
        if format[:1] == "'" or format[:1] == '"':
            return len(format)-1

        # address specifier
        two = format.split('&')
        if len(two) == 2:
            return self.calcPackSize(two[0], data)

        # code specifier
        two = format.split('=')
        if len(two) >= 2:
            return self.calcPackSize(two[0], data)

        # length specifier
        two = format.split('-')
        if len(two) == 2:
            return self.calcPackSize(two[0], data)

        # array specifier
        two = format.split('*')
        if len(two) == 2:
            answer = 0
            if two[0].isdigit():
                    if int(two[0]) != len(data):
                        raise Exception, "Array field has a constant size, and it doesn't match the actual value"
            elif two[0]:
                answer += self.calcPackSize(two[0], len(data))

            for each in data:
                answer += self.calcPackSize(two[1], each)
            return answer

        # "printf" string specifier
        if format[:1] == '%':
            # format string like specifier
            return len(format % data)

        # asciiz specifier
        if format[:1] == 'z':
            return len(data)+1

        # asciiz specifier
        if format[:1] == 'u':
            l = len(data)
            return l + (l & 1 and 3 or 2)

        # DCE-RPC/NDR string specifier
        if format[:1] == 'w':
            l = len(data)
            return 12+l+l % 2

        # literal specifier
        if format[:1] == ':':
            return len(data)

        # struct like specifier
        return calcsize(format)

    def calcUnpackSize(self, format, data, field = None):
        if self.debug:
            print "  calcUnpackSize( %s | %s | %r)" %  (field, format, data)

        # void specifier
        if format[:1] == '_':
            return 0

        addressField = self.findAddressFieldFor(field)
        if addressField is not None:
            if not self[addressField]:
                return 0

        try:
            lengthField = self.findLengthFieldFor(field)
            return self[lengthField]
        except:
            pass

        # XXX: Try to match to actual values, raise if no match
        
        # quote specifier
        if format[:1] == "'" or format[:1] == '"':
            return len(format)-1

        # address specifier
        two = format.split('&')
        if len(two) == 2:
            return self.calcUnpackSize(two[0], data)

        # code specifier
        two = format.split('=')
        if len(two) >= 2:
            return self.calcUnpackSize(two[0], data)

        # length specifier
        two = format.split('-')
        if len(two) == 2:
            return self.calcUnpackSize(two[0], data)

        # array specifier
        two = format.split('*')
        if len(two) == 2:
            answer = 0
            if two[0]:
                if two[0].isdigit():
                    number = int(two[0])
                else:
                    answer += self.calcUnpackSize(two[0], data)
                    number = self.unpack(two[0], data[:answer])

                while number:
                    number -= 1
                    answer += self.calcUnpackSize(two[1], data[answer:])
            else:
                while answer < len(data):
                    answer += self.calcUnpackSize(two[1], data[answer:])
            return answer

        # "printf" string specifier
        if format[:1] == '%':
            raise Exception, "Can't guess the size of a printf like specifier for unpacking"

        # asciiz specifier
        if format[:1] == 'z':
            return data.index('\x00')+1

        # asciiz specifier
        if format[:1] == 'u':
            l = data.index('\x00\x00')
            return l + (l & 1 and 3 or 2)

        # DCE-RPC/NDR string specifier
        if format[:1] == 'w':
            l = unpack('<L', data[:4])[0]
            return 12+l*2

        # literal specifier
        if format[:1] == ':':
            return len(data)

        # struct like specifier
        return calcsize(format)

    def calcPackFieldSize(self, fieldName, format = None):
        if format is None:
            format = self.formatForField(fieldName)

        return self.calcPackSize(format, self[fieldName])

    def formatForField(self, fieldName):
        for field in self.commonHdr+self.structure:
            if field[0] == fieldName:
                return field[1]
        raise Exception, ("Field %s not found" % fieldName)

    def findAddressFieldFor(self, fieldName):
        descriptor = '&%s' % fieldName
        l = len(descriptor)
        for field in self.commonHdr+self.structure:
            if field[1][-l:] == descriptor:
                return field[0]
        return None
        
    def findLengthFieldFor(self, fieldName):
        descriptor = '-%s' % fieldName
        l = len(descriptor)
        for field in self.commonHdr+self.structure:
            if field[1][-l:] == descriptor:
                return field[0]
        return None
        
    def zeroValue(self, format):
        two = format.split('*')
        if len(two) == 2:
            if two[0].isdigit():
                return (self.zeroValue(two[1]),)*int(two[0])
                        
        if not format.find('*') == -1: return ()
        if 's' in format: return ''
        if format in ['z',':','u']: return ''
        if format == 'w': return '\x00\x00'

        return 0

    def clear(self):
        for field in self.commonHdr + self.structure:
            self[field[0]] = self.zeroValue(field[1])

    def dump(self, msg = None, indent = 0):
        if msg is None: msg = self.__class__.__name__
        ind = ' '*indent
        print "\n%s" % msg
        fixedFields = []
        for field in self.commonHdr+self.structure:
            i = field[0] 
            if i in self.fields:
                fixedFields.append(i)
                if isinstance(self[i], Structure):
                    self[i].dump('%s%s:{' % (ind,i), indent = indent + 4)
                    print "%s}" % ind
                else:
                    print "%s%s: {%r}" % (ind,i,self[i])
        # Do we have remaining fields not defined in the structures? let's 
        # print them
        remainingFields = list(set(self.fields) - set(fixedFields))
        for i in remainingFields:
            if isinstance(self[i], Structure):
                self[i].dump('%s%s:{' % (ind,i), indent = indent + 4)
                print "%s}" % ind
            else:
                print "%s%s: {%r}" % (ind,i,self[i])


class _StructureTest:
    alignment = 0
    def create(self,data = None):
        if data is not None:
            return self.theClass(data, alignment = self.alignment)
        else:
            return self.theClass(alignment = self.alignment)

    def run(self):
        print
        print "-"*70
        testName = self.__class__.__name__
        print "starting test: %s....." % testName
        a = self.create()
        self.populate(a)
        a.dump("packing.....")
        a_str = str(a)
        print "packed: %r" % a_str
        print "unpacking....."
        b = self.create(a_str)
        b.dump("unpacked.....")
        print "repacking....."
        b_str = str(b)
        if b_str != a_str:
            print "ERROR: original packed and repacked don't match"
            print "packed: %r" % b_str

class _Test_simple(_StructureTest):
    class theClass(Structure):
        commonHdr = ()
        structure = (
                ('int1', '!L'),
                ('len1','!L-z1'),
                ('arr1','B*<L'),
                ('z1', 'z'),
                ('u1','u'),
                ('', '"COCA'),
                ('len2','!H-:1'),
                ('', '"COCA'),
                (':1', ':'),
                ('int3','>L'),
                ('code1','>L=len(arr1)*2+0x1000'),
                )

    def populate(self, a):
        a['default'] = 'hola'
        a['int1'] = 0x3131
        a['int3'] = 0x45444342
        a['z1']   = 'hola'
        a['u1']   = 'hola'.encode('utf_16_le')
        a[':1']   = ':1234:'
        a['arr1'] = (0x12341234,0x88990077,0x41414141)
        # a['len1'] = 0x42424242

class _Test_fixedLength(_Test_simple):
    def populate(self, a):
        _Test_simple.populate(self, a)
        a['len1'] = 0x42424242

class _Test_simple_aligned4(_Test_simple):
    alignment = 4

class _Test_nested(_StructureTest):
    class theClass(Structure):
        class _Inner(Structure):
            structure = (('data', 'z'),)

        structure = (
            ('nest1', ':', _Inner),
            ('nest2', ':', _Inner),
            ('int', '<L'),
        )

    def populate(self, a):
        a['nest1'] = _Test_nested.theClass._Inner()
        a['nest2'] = _Test_nested.theClass._Inner()
        a['nest1']['data'] = 'hola manola'
        a['nest2']['data'] = 'chau loco'
        a['int'] = 0x12345678
    
class _Test_Optional(_StructureTest):
    class theClass(Structure):
        structure = (
                ('pName','<L&Name'),
                ('pList','<L&List'),
                ('Name','w'),
                ('List','<H*<L'),
            )
            
    def populate(self, a):
        a['Name'] = 'Optional test'
        a['List'] = (1,2,3,4)
        
class _Test_Optional_sparse(_Test_Optional):
    def populate(self, a):
        _Test_Optional.populate(self, a)
        del a['Name']

class _Test_AsciiZArray(_StructureTest):
    class theClass(Structure):
        structure = (
            ('head','<L'),
            ('array','B*z'),
            ('tail','<L'),
        )

    def populate(self, a):
        a['head'] = 0x1234
        a['tail'] = 0xabcd
        a['array'] = ('hola','manola','te traje')
        
class _Test_UnpackCode(_StructureTest):
    class theClass(Structure):
        structure = (
            ('leni','<L=len(uno)*2'),
            ('cuchi','_-uno','leni/2'),
            ('uno',':'),
            ('dos',':'),
        )

    def populate(self, a):
        a['uno'] = 'soy un loco!'
        a['dos'] = 'que haces fiera'

class _Test_AAA(_StructureTest):
    class theClass(Structure):
        commonHdr = ()
        structure = (
          ('iv', '!L=((init_vector & 0xFFFFFF) << 8) | ((pad & 0x3f) << 2) | (keyid & 3)'),
          ('init_vector',   '_','(iv >> 8)'),
          ('pad',           '_','((iv >>2) & 0x3F)'),
          ('keyid',         '_','( iv & 0x03 )'),
          ('dataLen',       '_-data', 'len(inputDataLeft)-4'),
          ('data',':'),
          ('icv','>L'),
        )

    def populate(self, a):
        a['init_vector']=0x01020304
        #a['pad']=int('01010101',2)
        a['pad']=int('010101',2)
        a['keyid']=0x07
        a['data']="\xA0\xA1\xA2\xA3\xA4\xA5\xA6\xA7\xA8\xA9"
        a['icv'] = 0x05060708
        #a['iv'] = 0x01020304
        
if __name__ == '__main__':
    _Test_simple().run()

    try:
        _Test_fixedLength().run()
    except:
        print "cannot repack because length is bogus"

    _Test_simple_aligned4().run()
    _Test_nested().run()
    _Test_Optional().run()
    _Test_Optional_sparse().run()
    _Test_AsciiZArray().run()
    _Test_UnpackCode().run()
    _Test_AAA().run()