aboutsummaryrefslogtreecommitdiffstats
path: root/generator/nanopb_generator.py
blob: 9e85728641f607d9e72c991273d116e32f2c747e (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
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
#!/usr/bin/python

'''Generate header file for nanopb from a ProtoBuf FileDescriptorSet.'''
nanopb_version = "nanopb-0.2.4"

try:
    import google.protobuf.descriptor_pb2 as descriptor
    import google.protobuf.text_format as text_format
except:
    print
    print "*************************************************************"
    print "*** Could not import the Google protobuf Python libraries ***"
    print "*** Try installing package 'python-protobuf' or similar.  ***"
    print "*************************************************************"
    print
    raise

try:
    import nanopb_pb2
except:
    print
    print "***************************************************************"
    print "*** Could not import the precompiled nanopb_pb2.py.         ***"
    print "*** Run 'make' in the 'generator' folder to update the file.***"
    print "***************************************************************"
    print
    raise






# ---------------------------------------------------------------------------
#                     Generation of single fields
# ---------------------------------------------------------------------------

import time
import os.path

# Values are tuple (c type, pb type, encoded size)
FieldD = descriptor.FieldDescriptorProto
datatypes = {
    FieldD.TYPE_BOOL:       ('bool',     'BOOL',        1),
    FieldD.TYPE_DOUBLE:     ('double',   'DOUBLE',      8),
    FieldD.TYPE_FIXED32:    ('uint32_t', 'FIXED32',     4),
    FieldD.TYPE_FIXED64:    ('uint64_t', 'FIXED64',     8),
    FieldD.TYPE_FLOAT:      ('float',    'FLOAT',       4),
    FieldD.TYPE_INT32:      ('int32_t',  'INT32',       5),
    FieldD.TYPE_INT64:      ('int64_t',  'INT64',      10),
    FieldD.TYPE_SFIXED32:   ('int32_t',  'SFIXED32',    4),
    FieldD.TYPE_SFIXED64:   ('int64_t',  'SFIXED64',    8),
    FieldD.TYPE_SINT32:     ('int32_t',  'SINT32',      5),
    FieldD.TYPE_SINT64:     ('int64_t',  'SINT64',     10),
    FieldD.TYPE_UINT32:     ('uint32_t', 'UINT32',      5),
    FieldD.TYPE_UINT64:     ('uint64_t', 'UINT64',     10)
}

class Names:
    '''Keeps a set of nested names and formats them to C identifier.'''
    def __init__(self, parts = ()):
        if isinstance(parts, Names):
            parts = parts.parts
        self.parts = tuple(parts)
    
    def __str__(self):
        return '_'.join(self.parts)

    def __add__(self, other):
        if isinstance(other, (str, unicode)):
            return Names(self.parts + (other,))
        elif isinstance(other, tuple):
            return Names(self.parts + other)
        else:
            raise ValueError("Name parts should be of type str")
    
    def __eq__(self, other):
        return isinstance(other, Names) and self.parts == other.parts
    
def names_from_type_name(type_name):
    '''Parse Names() from FieldDescriptorProto type_name'''
    if type_name[0] != '.':
        raise NotImplementedError("Lookup of non-absolute type names is not supported")
    return Names(type_name[1:].split('.'))

def varint_max_size(max_value):
    '''Returns the maximum number of bytes a varint can take when encoded.'''
    for i in range(1, 11):
        if (max_value >> (i * 7)) == 0:
            return i
    raise ValueError("Value too large for varint: " + str(max_value))

assert varint_max_size(0) == 1
assert varint_max_size(127) == 1
assert varint_max_size(128) == 2

class EncodedSize:
    '''Class used to represent the encoded size of a field or a message.
    Consists of a combination of symbolic sizes and integer sizes.'''
    def __init__(self, value = 0, symbols = []):
        if isinstance(value, (str, Names)):
            symbols = [str(value)]
            value = 0
        self.value = value
        self.symbols = symbols
    
    def __add__(self, other):
        if isinstance(other, (int, long)):
            return EncodedSize(self.value + other, self.symbols)
        elif isinstance(other, (str, Names)):
            return EncodedSize(self.value, self.symbols + [str(other)])
        elif isinstance(other, EncodedSize):
            return EncodedSize(self.value + other.value, self.symbols + other.symbols)
        else:
            raise ValueError("Cannot add size: " + repr(other))

    def __mul__(self, other):
        if isinstance(other, (int, long)):
            return EncodedSize(self.value * other, [str(other) + '*' + s for s in self.symbols])
        else:
            raise ValueError("Cannot multiply size: " + repr(other))

    def __str__(self):
        if not self.symbols:
            return str(self.value)
        else:
            return '(' + str(self.value) + ' + ' + ' + '.join(self.symbols) + ')'

    def upperlimit(self):
        if not self.symbols:
            return self.value
        else:
            return 2**32 - 1

class Enum:
    def __init__(self, names, desc, enum_options):
        '''desc is EnumDescriptorProto'''
        
        self.options = enum_options
        self.names = names + desc.name
        
        if enum_options.long_names:
            self.values = [(self.names + x.name, x.number) for x in desc.value]            
        else:
            self.values = [(names + x.name, x.number) for x in desc.value] 
        
        self.value_longnames = [self.names + x.name for x in desc.value]
    
    def __str__(self):
        result = 'typedef enum _%s {\n' % self.names
        result += ',\n'.join(["    %s = %d" % x for x in self.values])
        result += '\n} %s;' % self.names
        return result

class Field:
    def __init__(self, struct_name, desc, field_options):
        '''desc is FieldDescriptorProto'''
        self.tag = desc.number
        self.struct_name = struct_name
        self.name = desc.name
        self.default = None
        self.max_size = None
        self.max_count = None
        self.array_decl = ""
        self.enc_size = None
        
        # Parse field options
        if field_options.HasField("max_size"):
            self.max_size = field_options.max_size
        
        if field_options.HasField("max_count"):
            self.max_count = field_options.max_count
        
        if desc.HasField('default_value'):
            self.default = desc.default_value
           
        # Check field rules, i.e. required/optional/repeated.
        can_be_static = True
        if desc.label == FieldD.LABEL_REQUIRED:
            self.rules = 'REQUIRED'
        elif desc.label == FieldD.LABEL_OPTIONAL:
            self.rules = 'OPTIONAL'
        elif desc.label == FieldD.LABEL_REPEATED:
            self.rules = 'REPEATED'
            if self.max_count is None:
                can_be_static = False
            else:
                self.array_decl = '[%d]' % self.max_count
        else:
            raise NotImplementedError(desc.label)
        
        # Decide the C data type to use in the struct.
        if datatypes.has_key(desc.type):
            self.ctype, self.pbtype, self.enc_size = datatypes[desc.type]
        elif desc.type == FieldD.TYPE_ENUM:
            self.pbtype = 'ENUM'
            self.ctype = names_from_type_name(desc.type_name)
            if self.default is not None:
                self.default = self.ctype + self.default
            self.enc_size = 5 # protoc rejects enum values > 32 bits
        elif desc.type == FieldD.TYPE_STRING:
            self.pbtype = 'STRING'
            if self.max_size is None:
                can_be_static = False
            else:
                self.ctype = 'char'
                self.array_decl += '[%d]' % self.max_size
                self.enc_size = varint_max_size(self.max_size) + self.max_size
        elif desc.type == FieldD.TYPE_BYTES:
            self.pbtype = 'BYTES'
            if self.max_size is None:
                can_be_static = False
            else:
                self.ctype = self.struct_name + self.name + 't'
                self.enc_size = varint_max_size(self.max_size) + self.max_size
        elif desc.type == FieldD.TYPE_MESSAGE:
            self.pbtype = 'MESSAGE'
            self.ctype = self.submsgname = names_from_type_name(desc.type_name)
            self.enc_size = None # Needs to be filled in after the message type is available
        else:
            raise NotImplementedError(desc.type)
        
        if field_options.type == nanopb_pb2.FT_DEFAULT:
            if can_be_static:
                field_options.type = nanopb_pb2.FT_STATIC
            else:
                field_options.type = nanopb_pb2.FT_CALLBACK
        
        if field_options.type == nanopb_pb2.FT_STATIC and not can_be_static:
            raise Exception("Field %s is defined as static, but max_size or max_count is not given." % self.name)
        
        if field_options.type == nanopb_pb2.FT_STATIC:
            self.allocation = 'STATIC'
        elif field_options.type == nanopb_pb2.FT_CALLBACK:
            self.allocation = 'CALLBACK'
            self.ctype = 'pb_callback_t'
            self.array_decl = ''
        else:
            raise NotImplementedError(field_options.type)
    
    def __cmp__(self, other):
        return cmp(self.tag, other.tag)
    
    def __str__(self):
        if self.rules == 'OPTIONAL' and self.allocation == 'STATIC':
            result = '    bool has_' + self.name + ';\n'
        elif self.rules == 'REPEATED' and self.allocation == 'STATIC':
            result = '    size_t ' + self.name + '_count;\n'
        else:
            result = ''
        result += '    %s %s%s;' % (self.ctype, self.name, self.array_decl)
        return result
    
    def types(self):
        '''Return definitions for any special types this field might need.'''
        if self.pbtype == 'BYTES' and self.allocation == 'STATIC':
            result = 'typedef struct {\n'
            result += '    size_t size;\n'
            result += '    uint8_t bytes[%d];\n' % self.max_size
            result += '} %s;\n' % self.ctype
        else:
            result = None
        return result
    
    def default_decl(self, declaration_only = False):
        '''Return definition for this field's default value.'''
        if self.default is None:
            return None

        ctype, default = self.ctype, self.default
        array_decl = ''
        
        if self.pbtype == 'STRING':
            if self.allocation != 'STATIC':
                return None # Not implemented
        
            array_decl = '[%d]' % self.max_size
            default = str(self.default).encode('string_escape')
            default = default.replace('"', '\\"')
            default = '"' + default + '"'
        elif self.pbtype == 'BYTES':
            if self.allocation != 'STATIC':
                return None # Not implemented

            data = self.default.decode('string_escape')
            data = ['0x%02x' % ord(c) for c in data]
            default = '{%d, {%s}}' % (len(data), ','.join(data))
        
        if declaration_only:
            return 'extern const %s %s_default%s;' % (ctype, self.struct_name + self.name, array_decl)
        else:
            return 'const %s %s_default%s = %s;' % (ctype, self.struct_name + self.name, array_decl, default)
    
    def tags(self):
        '''Return the #define for the tag number of this field.'''
        identifier = '%s_%s_tag' % (self.struct_name, self.name)
        return '#define %-40s %d\n' % (identifier, self.tag)
    
    def pb_field_t(self, prev_field_name):
        '''Return the pb_field_t initializer to use in the constant array.
        prev_field_name is the name of the previous field or None.
        '''
        result  = '    PB_FIELD2(%3d, ' % self.tag
        result += '%-8s, ' % self.pbtype
        result += '%s, ' % self.rules
        result += '%s, ' % self.allocation
        result += '%s, ' % ("FIRST" if not prev_field_name else "OTHER")
        result += '%s, ' % self.struct_name
        result += '%s, ' % self.name
        result += '%s, ' % (prev_field_name or self.name)
        
        if self.pbtype == 'MESSAGE':
            result += '&%s_fields)' % self.submsgname
        elif self.default is None:
            result += '0)'
        elif self.pbtype in ['BYTES', 'STRING'] and self.allocation != 'STATIC':
            result += '0)' # Arbitrary size default values not implemented
        else:
            result += '&%s_default)' % (self.struct_name + self.name)
        
        return result
    
    def largest_field_value(self):
        '''Determine if this field needs 16bit or 32bit pb_field_t structure to compile properly.
        Returns numeric value or a C-expression for assert.'''
        if self.pbtype == 'MESSAGE':
            if self.rules == 'REPEATED' and self.allocation == 'STATIC':
                return 'pb_membersize(%s, %s[0])' % (self.struct_name, self.name)
            else:
                return 'pb_membersize(%s, %s)' % (self.struct_name, self.name)

        return max(self.tag, self.max_size, self.max_count)        

    def encoded_size(self, allmsgs):
        '''Return the maximum size that this field can take when encoded,
        including the field tag. If the size cannot be determined, returns
        None.'''
        
        if self.allocation != 'STATIC':
            return None
        
        if self.pbtype == 'MESSAGE':
            for msg in allmsgs:
                if msg.name == self.submsgname:
                    encsize = msg.encoded_size(allmsgs)
                    if encsize is None:
                        return None # Submessage size is indeterminate
                        
                    # Include submessage length prefix
                    encsize += varint_max_size(encsize.upperlimit())
                    break
            else:
                # Submessage cannot be found, this currently occurs when
                # the submessage type is defined in a different file.
                # Instead of direct numeric value, reference the size that
                # has been #defined in the other file.
                encsize = EncodedSize(self.submsgname + 'size')

                # We will have to make a conservative assumption on the length
                # prefix size, though.
                encsize += 5

        elif self.enc_size is None:
            raise RuntimeError("Could not determine encoded size for %s.%s"
                               % (self.struct_name, self.name))
        else:
            encsize = EncodedSize(self.enc_size)
        
        encsize += varint_max_size(self.tag << 3) # Tag + wire type

        if self.rules == 'REPEATED':
            # Decoders must be always able to handle unpacked arrays.
            # Therefore we have to reserve space for it, even though
            # we emit packed arrays ourselves.
            encsize *= self.max_count
        
        return encsize


class ExtensionRange(Field):
    def __init__(self, struct_name, range_start, field_options):
        '''Implements a special pb_extension_t* field in an extensible message
        structure. The range_start signifies the index at which the extensions
        start. Not necessarily all tags above this are extensions, it is merely
        a speed optimization.
        '''
        self.tag = range_start
        self.struct_name = struct_name
        self.name = 'extensions'
        self.pbtype = 'EXTENSION'
        self.rules = 'OPTIONAL'
        self.allocation = 'CALLBACK'
        self.ctype = 'pb_extension_t'
        self.array_decl = ''
        self.default = None
        self.max_size = 0
        self.max_count = 0
        
    def __str__(self):
        return '    pb_extension_t *extensions;'
    
    def types(self):
        return None
    
    def tags(self):
        return ''
        
    def encoded_size(self, allmsgs):
        # We exclude extensions from the count, because they cannot be known
        # until runtime. Other option would be to return None here, but this
        # way the value remains useful if extensions are not used.
        return EncodedSize(0)

class ExtensionField(Field):
    def __init__(self, struct_name, desc, field_options):
        self.fullname = struct_name + desc.name
        self.extendee_name = names_from_type_name(desc.extendee)
        Field.__init__(self, self.fullname + 'struct', desc, field_options)
        
        if self.rules != 'OPTIONAL':
            self.skip = True
        else:
            self.skip = False
            self.rules = 'OPTEXT'

    def tags(self):
        '''Return the #define for the tag number of this field.'''
        identifier = '%s_tag' % self.fullname
        return '#define %-40s %d\n' % (identifier, self.tag)

    def extension_decl(self):
        '''Declaration of the extension type in the .pb.h file'''
        if self.skip:
            msg = '/* Extension field %s was skipped because only "optional"\n' % self.fullname
            msg +='   type of extension fields is currently supported. */\n'
            return msg
        
        return 'extern const pb_extension_type_t %s;\n' % self.fullname

    def extension_def(self):
        '''Definition of the extension type in the .pb.c file'''

        if self.skip:
            return ''

        result  = 'typedef struct {\n'
        result += str(self)
        result += '\n} %s;\n\n' % self.struct_name
        result += ('static const pb_field_t %s_field = \n  %s;\n\n' %
                    (self.fullname, self.pb_field_t(None)))
        result += 'const pb_extension_type_t %s = {\n' % self.fullname
        result += '    NULL,\n'
        result += '    NULL,\n'
        result += '    &%s_field\n' % self.fullname
        result += '};\n'
        return result


# ---------------------------------------------------------------------------
#                   Generation of messages (structures)
# ---------------------------------------------------------------------------


class Message:
    def __init__(self, names, desc, message_options):
        self.name = names
        self.fields = []
        
        for f in desc.field:
            field_options = get_nanopb_suboptions(f, message_options, self.name + f.name)
            if field_options.type != nanopb_pb2.FT_IGNORE:
                self.fields.append(Field(self.name, f, field_options))
        
        if len(desc.extension_range) > 0:
            field_options = get_nanopb_suboptions(desc, message_options, self.name + 'extensions')
            range_start = min([r.start for r in desc.extension_range])
            if field_options.type != nanopb_pb2.FT_IGNORE:
                self.fields.append(ExtensionRange(self.name, range_start, field_options))
        
        self.packed = message_options.packed_struct
        self.ordered_fields = self.fields[:]
        self.ordered_fields.sort()

    def get_dependencies(self):
        '''Get list of type names that this structure refers to.'''
        return [str(field.ctype) for field in self.fields]
    
    def __str__(self):
        result = 'typedef struct _%s {\n' % self.name

        if not self.ordered_fields:
            # Empty structs are not allowed in C standard.
            # Therefore add a dummy field if an empty message occurs.
            result += '    uint8_t dummy_field;'

        result += '\n'.join([str(f) for f in self.ordered_fields])
        result += '\n}'
        
        if self.packed:
            result += ' pb_packed'
        
        result += ' %s;' % self.name
        
        if self.packed:
            result = 'PB_PACKED_STRUCT_START\n' + result
            result += '\nPB_PACKED_STRUCT_END'
        
        return result
    
    def types(self):
        result = ""
        for field in self.fields:
            types = field.types()
            if types is not None:
                result += types + '\n'
        return result
    
    def default_decl(self, declaration_only = False):
        result = ""
        for field in self.fields:
            default = field.default_decl(declaration_only)
            if default is not None:
                result += default + '\n'
        return result

    def fields_declaration(self):
        result = 'extern const pb_field_t %s_fields[%d];' % (self.name, len(self.fields) + 1)
        return result

    def fields_definition(self):
        result = 'const pb_field_t %s_fields[%d] = {\n' % (self.name, len(self.fields) + 1)
        
        prev = None
        for field in self.ordered_fields:
            result += field.pb_field_t(prev)
            result += ',\n'
            prev = field.name
        
        result += '    PB_LAST_FIELD\n};'
        return result

    def encoded_size(self, allmsgs):
        '''Return the maximum size that this message can take when encoded.
        If the size cannot be determined, returns None.
        '''
        size = EncodedSize(0)
        for field in self.fields:
            fsize = field.encoded_size(allmsgs)
            if fsize is None:
                return None
            size += fsize
        
        return size


# ---------------------------------------------------------------------------
#                    Processing of entire .proto files
# ---------------------------------------------------------------------------


def iterate_messages(desc, names = Names()):
    '''Recursively find all messages. For each, yield name, DescriptorProto.'''
    if hasattr(desc, 'message_type'):
        submsgs = desc.message_type
    else:
        submsgs = desc.nested_type
    
    for submsg in submsgs:
        sub_names = names + submsg.name
        yield sub_names, submsg
        
        for x in iterate_messages(submsg, sub_names):
            yield x

def iterate_extensions(desc, names = Names()):
    '''Recursively find all extensions.
    For each, yield name, FieldDescriptorProto.
    '''
    for extension in desc.extension:
        yield names, extension

    for subname, subdesc in iterate_messages(desc, names):
        for extension in subdesc.extension:
            yield subname, extension

def parse_file(fdesc, file_options):
    '''Takes a FileDescriptorProto and returns tuple (enums, messages, extensions).'''
    
    enums = []
    messages = []
    extensions = []
    
    if fdesc.package:
        base_name = Names(fdesc.package.split('.'))
    else:
        base_name = Names()
    
    for enum in fdesc.enum_type:
        enum_options = get_nanopb_suboptions(enum, file_options, base_name + enum.name)
        enums.append(Enum(base_name, enum, enum_options))
    
    for names, message in iterate_messages(fdesc, base_name):
        message_options = get_nanopb_suboptions(message, file_options, names)
        messages.append(Message(names, message, message_options))
        for enum in message.enum_type:
            enum_options = get_nanopb_suboptions(enum, message_options, names + enum.name)
            enums.append(Enum(names, enum, enum_options))
    
    for names, extension in iterate_extensions(fdesc, base_name):
        field_options = get_nanopb_suboptions(extension, file_options, names)
        if field_options.type != nanopb_pb2.FT_IGNORE:
            extensions.append(ExtensionField(names, extension, field_options))
    
    # Fix field default values where enum short names are used.
    for enum in enums:
        if not enum.options.long_names:
            for message in messages:
                for field in message.fields:
                    if field.default in enum.value_longnames:
                        idx = enum.value_longnames.index(field.default)
                        field.default = enum.values[idx][0]
    
    return enums, messages, extensions

def toposort2(data):
    '''Topological sort.
    From http://code.activestate.com/recipes/577413-topological-sort/
    This function is under the MIT license.
    '''
    for k, v in data.items():
        v.discard(k) # Ignore self dependencies
    extra_items_in_deps = reduce(set.union, data.values(), set()) - set(data.keys())
    data.update(dict([(item, set()) for item in extra_items_in_deps]))
    while True:
        ordered = set(item for item,dep in data.items() if not dep)
        if not ordered:
            break
        for item in sorted(ordered):
            yield item
        data = dict([(item, (dep - ordered)) for item,dep in data.items()
                if item not in ordered])
    assert not data, "A cyclic dependency exists amongst %r" % data

def sort_dependencies(messages):
    '''Sort a list of Messages based on dependencies.'''
    dependencies = {}
    message_by_name = {}
    for message in messages:
        dependencies[str(message.name)] = set(message.get_dependencies())
        message_by_name[str(message.name)] = message
    
    for msgname in toposort2(dependencies):
        if msgname in message_by_name:
            yield message_by_name[msgname]

def make_identifier(headername):
    '''Make #ifndef identifier that contains uppercase A-Z and digits 0-9'''
    result = ""
    for c in headername.upper():
        if c.isalnum():
            result += c
        else:
            result += '_'
    return result

def generate_header(dependencies, headername, enums, messages, extensions, options):
    '''Generate content for a header file.
    Generates strings, which should be concatenated and stored to file.
    '''
    
    yield '/* Automatically generated nanopb header */\n'
    yield '/* Generated by %s at %s. */\n\n' % (nanopb_version, time.asctime())
    
    symbol = make_identifier(headername)
    yield '#ifndef _PB_%s_\n' % symbol
    yield '#define _PB_%s_\n' % symbol
    try:
        yield options.libformat % ('pb.h')
    except TypeError:
        # no %s specified - use whatever was passed in as options.libformat
        yield options.libformat
    yield '\n'
    
    for dependency in dependencies:
        noext = os.path.splitext(dependency)[0]
        yield options.genformat % (noext + '.' + options.extension + '.h')
        yield '\n'

    yield '#ifdef __cplusplus\n'
    yield 'extern "C" {\n'
    yield '#endif\n\n'
    
    yield '/* Enum definitions */\n'
    for enum in enums:
        yield str(enum) + '\n\n'
    
    yield '/* Struct definitions */\n'
    for msg in sort_dependencies(messages):
        yield msg.types()
        yield str(msg) + '\n\n'
    
    if extensions:
        yield '/* Extensions */\n'
        for extension in extensions:
            yield extension.extension_decl()
        yield '\n'
        
    yield '/* Default values for struct fields */\n'
    for msg in messages:
        yield msg.default_decl(True)
    yield '\n'
    
    yield '/* Field tags (for use in manual encoding/decoding) */\n'
    for msg in sort_dependencies(messages):
        for field in msg.fields:
            yield field.tags()
    for extension in extensions:
        yield extension.tags()
    yield '\n'
    
    yield '/* Struct field encoding specification for nanopb */\n'
    for msg in messages:
        yield msg.fields_declaration() + '\n'
    yield '\n'
    
    yield '/* Maximum encoded size of messages (where known) */\n'
    for msg in messages:
        msize = msg.encoded_size(messages)
        if msize is not None:
            identifier = '%s_size' % msg.name
            yield '#define %-40s %s\n' % (identifier, msize)
    yield '\n'
    
    yield '#ifdef __cplusplus\n'
    yield '} /* extern "C" */\n'
    yield '#endif\n'
    
    # End of header
    yield '\n#endif\n'

def generate_source(headername, enums, messages, extensions, options):
    '''Generate content for a source file.'''
    
    yield '/* Automatically generated nanopb constant definitions */\n'
    yield '/* Generated by %s at %s. */\n\n' % (nanopb_version, time.asctime())
    yield options.genformat % (headername)
    yield '\n'
    
    for msg in messages:
        yield msg.default_decl(False)
    
    yield '\n\n'
    
    for msg in messages:
        yield msg.fields_definition() + '\n\n'
    
    for ext in extensions:
        yield ext.extension_def() + '\n'
        
    # Add checks for numeric limits
    if messages:
        count_required_fields = lambda m: len([f for f in msg.fields if f.rules == 'REQUIRED'])
        largest_msg = max(messages, key = count_required_fields)
        largest_count = count_required_fields(largest_msg)
        if largest_count > 64:
            yield '\n/* Check that missing required fields will be properly detected */\n'
            yield '#if PB_MAX_REQUIRED_FIELDS < %d\n' % largest_count
            yield '#error Properly detecting missing required fields in %s requires \\\n' % largest_msg.name
            yield '       setting PB_MAX_REQUIRED_FIELDS to %d or more.\n' % largest_count
            yield '#endif\n'
    
    worst = 0
    worst_field = ''
    checks = []
    checks_msgnames = []
    for msg in messages:
        checks_msgnames.append(msg.name)
        for field in msg.fields:
            status = field.largest_field_value()
            if isinstance(status, (str, unicode)):
                checks.append(status)
            elif status > worst:
                worst = status
                worst_field = str(field.struct_name) + '.' + str(field.name)

    if worst > 255 or checks:
        yield '\n/* Check that field information fits in pb_field_t */\n'
        
        if worst < 65536:
            yield '#if !defined(PB_FIELD_16BIT) && !defined(PB_FIELD_32BIT)\n'
            if worst > 255:
                yield '#error Field descriptor for %s is too large. Define PB_FIELD_16BIT to fix this.\n' % worst_field
            else:
                assertion = ' && '.join(str(c) + ' < 256' for c in checks)
                msgs = '_'.join(str(n) for n in checks_msgnames)
                yield 'STATIC_ASSERT((%s), YOU_MUST_DEFINE_PB_FIELD_16BIT_FOR_MESSAGES_%s)\n'%(assertion,msgs)
            yield '#endif\n\n'
        
        if worst > 65535 or checks:
            yield '#if !defined(PB_FIELD_32BIT)\n'
            if worst > 65535:
                yield '#error Field descriptor for %s is too large. Define PB_FIELD_32BIT to fix this.\n' % worst_field
            else:
                assertion = ' && '.join(str(c) + ' < 65536' for c in checks)
                msgs = '_'.join(str(n) for n in checks_msgnames)
                yield 'STATIC_ASSERT((%s), YOU_MUST_DEFINE_PB_FIELD_32BIT_FOR_MESSAGES_%s)\n'%(assertion,msgs)
            yield '#endif\n'
    
    # Add check for sizeof(double)
    has_double = False
    for msg in messages:
        for field in msg.fields:
            if field.ctype == 'double':
                has_double = True
    
    if has_double:
        yield '\n'
        yield '/* On some platforms (such as AVR), double is really float.\n'
        yield ' * These are not directly supported by nanopb, but see example_avr_double.\n'
        yield ' * To get rid of this error, remove any double fields from your .proto.\n'
        yield ' */\n'
        yield 'STATIC_ASSERT(sizeof(double) == 8, DOUBLE_MUST_BE_8_BYTES)\n'
    
    yield '\n'

# ---------------------------------------------------------------------------
#                    Options parsing for the .proto files
# ---------------------------------------------------------------------------

from fnmatch import fnmatch

def read_options_file(infile):
    '''Parse a separate options file to list:
        [(namemask, options), ...]
    '''
    results = []
    for line in infile:
        line = line.strip()
        if not line or line.startswith('//') or line.startswith('#'):
            continue
        
        parts = line.split(None, 1)
        opts = nanopb_pb2.NanoPBOptions()
        text_format.Merge(parts[1], opts)
        results.append((parts[0], opts))

    return results

class Globals:
    '''Ugly global variables, should find a good way to pass these.'''
    verbose_options = False
    separate_options = []

def get_nanopb_suboptions(subdesc, options, name):
    '''Get copy of options, and merge information from subdesc.'''
    new_options = nanopb_pb2.NanoPBOptions()
    new_options.CopyFrom(options)
    
    # Handle options defined in a separate file
    dotname = '.'.join(name.parts)
    for namemask, options in Globals.separate_options:
        if fnmatch(dotname, namemask):
            new_options.MergeFrom(options)
    
    # Handle options defined in .proto
    if isinstance(subdesc.options, descriptor.FieldOptions):
        ext_type = nanopb_pb2.nanopb
    elif isinstance(subdesc.options, descriptor.FileOptions):
        ext_type = nanopb_pb2.nanopb_fileopt
    elif isinstance(subdesc.options, descriptor.MessageOptions):
        ext_type = nanopb_pb2.nanopb_msgopt
    elif isinstance(subdesc.options, descriptor.EnumOptions):
        ext_type = nanopb_pb2.nanopb_enumopt
    else:
        raise Exception("Unknown options type")
    
    if subdesc.options.HasExtension(ext_type):
        ext = subdesc.options.Extensions[ext_type]
        new_options.MergeFrom(ext)
    
    if Globals.verbose_options:
        print "Options for " + dotname + ":"
        print text_format.MessageToString(new_options)
    
    return new_options


# ---------------------------------------------------------------------------
#                         Command line interface
# ---------------------------------------------------------------------------

import sys
import os.path    
from optparse import OptionParser

optparser = OptionParser(
    usage = "Usage: nanopb_generator.py [options] file.pb ...",
    epilog = "Compile file.pb from file.proto by: 'protoc -ofile.pb file.proto'. " +
             "Output will be written to file.pb.h and file.pb.c.")
optparser.add_option("-x", dest="exclude", metavar="FILE", action="append", default=[],
    help="Exclude file from generated #include list.")
optparser.add_option("-e", "--extension", dest="extension", metavar="EXTENSION", default="pb",
    help="Set extension to use instead of 'pb' for generated files. [default: %default]")
optparser.add_option("-f", "--options-file", dest="options_file", metavar="FILE", default="%s.options",
    help="Set name of a separate generator options file.")
optparser.add_option("-Q", "--generated-include-format", dest="genformat",
    metavar="FORMAT", default='#include "%s"\n',
    help="Set format string to use for including other .pb.h files. [default: %default]")
optparser.add_option("-L", "--library-include-format", dest="libformat",
    metavar="FORMAT", default='#include <%s>\n',
    help="Set format string to use for including the nanopb pb.h header. [default: %default]")
optparser.add_option("-q", "--quiet", dest="quiet", action="store_true", default=False,
    help="Don't print anything except errors.")
optparser.add_option("-v", "--verbose", dest="verbose", action="store_true", default=False,
    help="Print more information.")
optparser.add_option("-s", dest="settings", metavar="OPTION:VALUE", action="append", default=[],
    help="Set generator option (max_size, max_count etc.).")

def process_file(filename, fdesc, options):
    '''Process a single file.
    filename: The full path to the .proto or .pb source file, as string.
    fdesc: The loaded FileDescriptorSet, or None to read from the input file.
    options: Command line options as they come from OptionsParser.
    
    Returns a dict:
        {'headername': Name of header file,
         'headerdata': Data for the .h header file,
         'sourcename': Name of the source code file,
         'sourcedata': Data for the .c source code file
        }
    '''
    toplevel_options = nanopb_pb2.NanoPBOptions()
    for s in options.settings:
        text_format.Merge(s, toplevel_options)
    
    if not fdesc:
        data = open(filename, 'rb').read()
        fdesc = descriptor.FileDescriptorSet.FromString(data).file[0]
    
    # Check if there is a separate .options file
    try:
        optfilename = options.options_file % os.path.splitext(filename)[0]
    except TypeError:
        # No %s specified, use the filename as-is
        optfilename = options.options_file
    
    if options.verbose:
        print 'Reading options from ' + optfilename
    
    if os.path.isfile(optfilename):
        Globals.separate_options = read_options_file(open(optfilename, "rU"))
    else:
        Globals.separate_options = []
    
    # Parse the file
    file_options = get_nanopb_suboptions(fdesc, toplevel_options, Names([filename]))
    enums, messages, extensions = parse_file(fdesc, file_options)
    
    # Decide the file names
    noext = os.path.splitext(filename)[0]
    headername = noext + '.' + options.extension + '.h'
    sourcename = noext + '.' + options.extension + '.c'
    headerbasename = os.path.basename(headername)
    
    # List of .proto files that should not be included in the C header file
    # even if they are mentioned in the source .proto.
    excludes = ['nanopb.proto', 'google/protobuf/descriptor.proto'] + options.exclude
    dependencies = [d for d in fdesc.dependency if d not in excludes]
    
    headerdata = ''.join(generate_header(dependencies, headerbasename, enums,
                                         messages, extensions, options))

    sourcedata = ''.join(generate_source(headerbasename, enums,
                                         messages, extensions, options))

    return {'headername': headername, 'headerdata': headerdata,
            'sourcename': sourcename, 'sourcedata': sourcedata}
    
def main_cli():
    '''Main function when invoked directly from the command line.'''
    
    options, filenames = optparser.parse_args()
    
    if not filenames:
        optparser.print_help()
        sys.exit(1)
    
    if options.quiet:
        options.verbose = False

    Globals.verbose_options = options.verbose
    
    for filename in filenames:
        results = process_file(filename, None, options)
        
        if not options.quiet:
            print "Writing to " + results['headername'] + " and " + results['sourcename']
    
        open(results['headername'], 'w').write(results['headerdata'])
        open(results['sourcename'], 'w').write(results['sourcedata'])        

def main_plugin():
    '''Main function when invoked as a protoc plugin.'''

    import plugin_pb2
    data = sys.stdin.read()
    request = plugin_pb2.CodeGeneratorRequest.FromString(data)
    
    import shlex
    args = shlex.split(request.parameter)
    options, dummy = optparser.parse_args(args)
    
    # We can't go printing stuff to stdout
    Globals.verbose_options = False
    options.verbose = False
    options.quiet = True
    
    response = plugin_pb2.CodeGeneratorResponse()
    
    for filename in request.file_to_generate:
        for fdesc in request.proto_file:
            if fdesc.name == filename:
                results = process_file(filename, fdesc, options)
                
                f = response.file.add()
                f.name = results['headername']
                f.content = results['headerdata']

                f = response.file.add()
                f.name = results['sourcename']
                f.content = results['sourcedata']    
    
    sys.stdout.write(response.SerializeToString())

if __name__ == '__main__':
    # Check if we are running as a plugin under protoc
    if 'protoc-gen-' in sys.argv[0]:
        main_plugin()
    else:
        main_cli()