Get rid of pb_bytes_ptr_t, just allocate pb_bytes_array_t dynamically.

This makes the internal logic much simpler, and also keeps the datatypes
more similar between STATIC/POINTER cases. It will still be a bit cumbersome
to use because of variable length array member. Macros PB_BYTES_ARRAY_T(n) and
PB_BYTES_ARRAY_T_ALLOCSIZE(n) have been added to make life a bit easier.

This has the drawback that it is no longer as easy to use externally allocated
byte array as input for bytes field in pointer mode. However, this is still
easy to do using callbacks, so it shouldn't be a large issue.

generator/nanopb_generator.py

@@ -246,7 +246,7 @@ class Field:
                 self.ctype = self.struct_name + self.name + 't'
                 self.enc_size = varint_max_size(self.max_size) + self.max_size
             elif self.allocation == 'POINTER':
-                self.ctype = 'pb_bytes_ptr_t'
+                self.ctype = 'pb_bytes_array_t'
         elif desc.type == FieldD.TYPE_MESSAGE:
             self.pbtype = 'MESSAGE'
             self.ctype = self.submsgname = names_from_type_name(desc.type_name)
@@ -266,8 +266,8 @@ class Field:
             if self.pbtype == 'MESSAGE':
                 # Use struct definition, so recursive submessages are possible
                 result += '    struct _%s *%s;' % (self.ctype, self.name)
-            elif self.rules == 'REPEATED' and self.pbtype == 'STRING':
-                # String arrays need to be defined as pointers to pointers
+            elif self.rules == 'REPEATED' and self.pbtype in ['STRING', 'BYTES']:
+                # String/bytes arrays need to be defined as pointers to pointers
                 result += '    %s **%s;' % (self.ctype, self.name)
             else:
                 result += '    %s *%s;' % (self.ctype, self.name)