Added note in docs that this is the dev branch.

Conflicts:
	tests/Makefile
	tests/test_encode1.c

LICENSE

@@ -1,4 +1,5 @@
 Copyright (c) 2011 Petteri Aimonen <jpa at nanopb.mail.kapsi.fi>
+Copyright (c) 2011 Michael Poole <mdpoole@troilus.org>
 
 This software is provided 'as-is', without any express or 
 implied warranty. In no event will the authors be held liable 

docs/concepts.rst

@@ -74,7 +74,7 @@ This is the way to get the size of the message without storing it anywhere::
 
  Person myperson = ...;
  pb_ostream_t sizestream = {0};
- pb_encode(&sizestream, Person_fields, &myperson);
+ pb_encode(&sizestream, Person_msg, &myperson);
  printf("Encoded size is %d\n", sizestream.bytes_written);
 
 **Example 2:**
@@ -142,6 +142,7 @@ Most Protocol Buffers datatypes have directly corresponding C datatypes, such as
 1) Strings, bytes and repeated fields of any type map to callback functions by default.
 2) If there is a special option *(nanopb).max_size* specified in the .proto file, string maps to null-terminated char array and bytes map to a structure containing a char array and a size field.
 3) If there is a special option *(nanopb).max_count* specified on a repeated field, it maps to an array of whatever type is being repeated. Another field will be created for the actual number of entries stored.
+4) The option *(nanopb).pointer* can override the default (false, unless the *-p* option is passed to *nanopb_generator.py*) behavior.  If a string, byte, or submessage is generated as a pointer field, and it is repeated (with a maximum count), required, or optional, the members will be pointers rather than in-line data.
 
 =============================================================================== =======================
       field in .proto                                                           autogenerated in .h
@@ -156,10 +157,19 @@ required bytes data = 1 [(nanopb).max_size = 40];
                                                                                 |    uint8_t bytes[40];
                                                                                 | } Person_data_t;
                                                                                 | Person_data_t data;
+required string name = 1 [(nanopb).pointer = true];                             char \*name;
+required bytes data = 1 [(nanopb).pointer = true];                              pb_bytes_t data;
+required bytes data = 1 [(nanopb).pointer = true, (nanopb).max_count = 5];      | size_t data_count;
+                                                                                | pb_bytes_t data[5];
+required Message msg = 1 [(nanopb).pointer = true];                             Message \*msg;
 =============================================================================== =======================
 
 The maximum lengths are checked in runtime. If string/bytes/array exceeds the allocated length, *pb_decode* will return false. 
 
+If a pointer-type field is not received, the field will be marked as absent, but the pointer will not be modified.  This helps reduce memory fragmentation and churn, but increases worst-case memory usage and means you must use the *Message_has(object, Field)* macro rather than testing for a null pointer.  You can use the `pb_clean`_ function to release unused memory in these cases.
+
+.. _`pb_clean`: reference.html#pb-clean
+
 Field callbacks
 ===============
 When a field has dynamic length, nanopb cannot statically allocate storage for it. Instead, it allows you to handle the field in whatever way you want, using a callback function.
@@ -173,16 +183,20 @@ Encoding callbacks
 ::
 
     bool (*encode)(pb_ostream_t *stream, const pb_field_t *field, const void *arg);
+    bool (*encode_buffer)(pb_strstream_t *stream, const pb_field_t *field, const void *arg);
 
-When encoding, the callback should write out complete fields, including the wire type and field number tag. It can write as many or as few fields as it likes. For example, if you want to write out an array as *repeated* field, you should do it all in a single call.
+When encoding, the callbacks should write out complete fields, including the wire type and field number tag. The callback can write as many or as few fields as it likes. For example, if you want to write out an array as *repeated* field, you should do it all in a single call.
 
-Usually you can use `pb_encode_tag_for_field`_ to encode the wire type and tag number of the field. However, if you want to encode a repeated field as a packed array, you must call `pb_encode_tag`_ instead to specify a wire type of *PB_WT_STRING*.
+Usually you can use `pb_encode_tag_for_field`_ (or `pb_encbuf_tag_for_field`_ for the *encode_buffer* callback) to encode the wire type and tag number of the field. However, if you want to encode a repeated field as a packed array, you must call `pb_encode_tag`_ (respectively, `pb_encbuf_tag`_) instead to specify a wire type of *PB_WT_STRING*.
 
-If the callback is used in a submessage, it will be called multiple times during a single call to `pb_encode`_. In this case, it must produce the same amount of data every time. If the callback is directly in the main message, it is called only once.
+If the callback is used in a submessage, *encode* will be called multiple times during a single call to `pb_encode`_. In this case, it must produce the same amount of data every time. If the callback is directly in the main message, or if you are using `pb_encode_buffer`_, the callback is called only once.
 
 .. _`pb_encode`: reference.html#pb-encode
+.. _`pb_encode_buffer`: reference.html#pb-encode-buffer
 .. _`pb_encode_tag_for_field`: reference.html#pb-encode-tag-for-field
+.. _`pb_encbuf_tag_for_field`: reference.html#pb-encbuf-tag-for-field
 .. _`pb_encode_tag`: reference.html#pb-encode-tag
+.. _`pb_encbuf_tag`: reference.html#pb-encbuf-tag
 
 This callback writes out a dynamically sized string::
 
@@ -195,6 +209,17 @@ This callback writes out a dynamically sized string::
         return pb_encode_string(stream, (uint8_t*)str, strlen(str));
     }
 
+The equivalent for in-memory buffers has to write the elements in the opposite order, because the buffer writers prepend their data::
+
+    bool write_string_buf(pb_strstream_t *stream, const pb_field_t *field, const void *arg)
+    {
+        char *str = get_string_from_somewhere();
+        if (!pb_encbuf_string(stream, (uint8_t*)str, strlen(str)))
+            return false;
+        
+        return pb_encbuf_tag_for_field(stream, field));
+    }
+
 Decoding callbacks
 ------------------
 ::
@@ -221,10 +246,10 @@ This callback reads multiple integers and prints them::
         return true;
     }
 
-Field description array
-=======================
+Message descriptor
+==================
 
-For using the *pb_encode* and *pb_decode* functions, you need an array of pb_field_t constants describing the structure you wish to encode. This description is usually autogenerated from .proto file.
+For using the *pb_encode*, *pb_encode_buffer* and *pb_decode* functions, you need a message descriptor describing the structure you wish to encode. This description is usually autogenerated from .proto file.
 
 For example this submessage in the Person.proto file::
 
@@ -235,22 +260,51 @@ For example this submessage in the Person.proto file::
     }
  }
 
-generates this field description array for the structure *Person_PhoneNumber*::
+generates these declarations and definitions for the structure *Person_PhoneNumber*::
+
+ typedef PB_MSG_STRUCT(2) Person_PhoneNumber_msg_t;
+ extern const Person_PhoneNumber_msg_t Person_PhoneNumber_real_msg;
+ #define Person_PhoneNumber_msg ((const pb_message_t*)&Person_PhoneNumber_real_msg)
+
+ const Person_PhoneNumber_msg_t Person_PhoneNumber_real_msg = {
+    2,
+    {
 
- const pb_field_t Person_PhoneNumber_fields[3] = {
     {1, PB_HTYPE_REQUIRED | PB_LTYPE_STRING,
     offsetof(Person_PhoneNumber, number), 0,
     pb_membersize(Person_PhoneNumber, number), 0, 0},
 
     {2, PB_HTYPE_OPTIONAL | PB_LTYPE_VARINT,
-    pb_delta(Person_PhoneNumber, type, number),
-    pb_delta(Person_PhoneNumber, has_type, type),
+    pb_delta_end(Person_PhoneNumber, type, number), 0,
     pb_membersize(Person_PhoneNumber, type), 0,
     &Person_PhoneNumber_type_default},
 
-    PB_LAST_FIELD
+    }
  };
 
+ #define Person_PhoneNumber_has(STRUCT, FIELD) PB_HAS_FIELD(STRUCT, Person_PhoneNumber, FIELD)
+ #define Person_PhoneNumber_set(STRUCT, FIELD) PB_SET_FIELD(STRUCT, Person_PhoneNumber, FIELD)
+ #define Person_PhoneNumber_clear(STRUCT, FIELD) PB_CLEAR_FIELD(STRUCT, Person_PhoneNumber, FIELD)
+ #define Person_PhoneNumber_number_index 0
+ #define Person_PhoneNumber_number_tag 1
+ #define Person_PhoneNumber_type_index 1
+ #define Person_PhoneNumber_type_tag 2
+
+Optional Fields
+===============
+
+The *has_fields* member of each generated structure is an array where
+each bit indicates the presence of the corresponding (optional) field.
+The generated header file provides helper macros to read and update
+that array; in the previous example, they are
+*Person_PhoneNumber_has*, *Person_PhoneNumber_set* and
+*Person_PhoneNumber_clear*.
+
+For convenience, *pb_encode* and *pb_encode_buffer* only check these
+bits for optional fields.  *pb_decode* sets the corresponding bit for
+every field it decodes, whether the field is optional or not.
+
+.. Should there be a section here on pointer fields?
 
 Return values and error handling
 ================================

docs/index.rst

@@ -14,6 +14,9 @@ Overall structure
 
 For the runtime program, you always need *pb.h* for type declarations.
 Depending on whether you want to encode, decode, or both, you also need *pb_encode.h/c* or *pb_decode.h/c*.
+If you only encode into in-memory buffers, *pb_decode_buffer.h/c* should be slightly faster and smaller.
+
+If your *.proto* file encodes submessages or other fields using pointers, you must compile *pb_decode.c* with a preprocessor macro named *MALLOC_HEADER* that is the name of a header with definitions (either as functions or macros) for *calloc()*, *realloc()* and *free()*.  For a typical hosted configuration, this should be *<stdlib.h>*.
 
 The high-level encoding and decoding functions take an array of *pb_field_t* structures, which describes the fields of a message structure. Usually you want these autogenerated from a *.proto* file. The tool script *nanopb_generator.py* accomplishes this.
 
@@ -25,9 +28,10 @@ So a typical project might include these files:
     - pb.h
     - pb_decode.h and pb_decode.c (needed for decoding messages)
     - pb_encode.h and pb_encode.c (needed for encoding messages)
+    - pb_encode_buffer.h and pb_encode_buffer.c (for encoding specifically into in-memory buffers)
 2) Protocol description (you can have many):
     - person.proto (just an example)
-    - person.pb.c (autogenerated, contains initializers for const arrays)
+    - person.pb.c (autogenerated, contains initializers for message descriptors)
     - person.pb.h (autogenerated, contains type declarations)
 
 Features and limitations
@@ -52,7 +56,7 @@ Features and limitations
 #) Unknown fields are not preserved when decoding and re-encoding a message.
 #) Reflection (runtime introspection) is not supported. E.g. you can't request a field by giving its name in a string.
 #) Numeric arrays are always encoded as packed, even if not marked as packed in .proto. This causes incompatibility with decoders that do not support packed format.
-#) Cyclic references between messages are not supported. They could be supported in callback-mode if there was an option in the generator to set the mode.
+#) Limited support for cyclic references between messages.  (The cycle must be broken by making one of the references a pointer or callback field, and objects that have circular references are not detected when encoding.)
 
 Getting started
 ===============
@@ -74,17 +78,30 @@ You should now have in *message.pb.h*::
     int32_t value;
  } Example;
  
- extern const pb_field_t Example_fields[2];
+ typedef PB_MSG_STRUCT(1) Example_msg_t;
+ extern const Example_msg_t Example_real_msg;
+ #define Example_msg ((const pb_message_t*)&Example_real_msg)
 
 Now in your main program do this to encode a message::
 
  Example mymessage = {42};
  uint8_t buffer[10];
  pb_ostream_t stream = pb_ostream_from_buffer(buffer, sizeof(buffer));
- pb_encode(&stream, Example_fields, &mymessage);
+ pb_encode(&stream, Example_msg, &mymessage);
 
 After that, buffer will contain the encoded message.
 The number of bytes in the message is stored in *stream.bytes_written*.
+
+Using *pb_encode_buffer.h/c* interface is very similar::
+
+ Example mymessage = {42};
+ uint8_t buffer[10];
+ pb_strstream_t stream = pb_str_from_buffer(buffer, sizeof(buffer));
+ pb_encode_buffer(&stream, Example_msg, &mymessage);
+
+The encoded message will start at *stream.last* and continue until the
+end of *buffer* (that is, it has length *buffer - stream.last*).
+
 You can feed the message to *protoc --decode=Example message.proto* to verify its validity.
 
 For complete examples of the simple cases, see *tests/test_decode1.c* and *tests/test_encode1.c*. For an example with network interface, see the *example* subdirectory.
@@ -108,6 +125,5 @@ This also generates a file called *breakpoints* which includes all lines returni
 
 Wishlist
 ========
-#) A specialized encoder for encoding to a memory buffer. Should serialize in reverse order to avoid having to determine submessage size beforehand.
 #) A cleaner rewrite of the Python-based source generator.
 #) Better performance for 16- and 8-bit platforms: use smaller datatypes where possible.

docs/menu.rst

@@ -3,7 +3,11 @@
     1) `Overview`_
     2) `Concepts`_
     3) `API reference`_
-    
+
+Note: This is the documentation for the *dynamic_alloc_dev* development branch. See here for the 
+documentation for the latest stable version:
+http://koti.kapsi.fi/~jpa/nanopb/docs/index.html
+
 .. _`Overview`: index.html
 .. _`Concepts`: concepts.html
 .. _`API reference`: reference.html

docs/reference.rst

@@ -13,9 +13,9 @@ pb_type_t
 ---------
 Defines the encoder/decoder behaviour that should be used for a field. ::
 
-    typedef enum { ... } pb_type_t;
+    typedef uint8_t pb_type_t;
 
-The low-order byte of the enumeration values defines the function that can be used for encoding and decoding the field data:
+The low-order nibble of the enumeration values defines the function that can be used for encoding and decoding the field data:
 
 ==================== ===== ================================================
 LTYPE identifier     Value Storage format
@@ -28,19 +28,23 @@ PB_LTYPE_STRING      0x04  Null-terminated string.
 PB_LTYPE_SUBMESSAGE  0x05  Submessage structure.
 ==================== ===== ================================================
 
-The high-order byte defines whether the field is required, optional, repeated or callback:
+The high-order nibble defines whether the field is required, optional, repeated or callback, and whether it is a pointer:
 
 ==================== ===== ================================================
 HTYPE identifier     Value Field handling
 ==================== ===== ================================================
 PB_HTYPE_REQUIRED    0x00  Verify that field exists in decoded message.
-PB_HTYPE_OPTIONAL    0x10  Use separate *has_<field>* boolean to specify
-                           whether the field is present.
+PB_HTYPE_OPTIONAL    0x10  Use the structure's *has_fields* bit array to
+                           specify whether the field is present.
 PB_HTYPE_ARRAY       0x20  A repeated field with preallocated array.
                            Separate *<field>_count* for number of items.
 PB_HTYPE_CALLBACK    0x30  A field with dynamic storage size, data is
                            actually a pointer to a structure containing a
                            callback function.
+PB_HTYPE_POINTER     0x80  For required, optional and array non-scalar
+                           fields, uses a pointer type (char* for strings,
+                           pb_bytes_t for bytes, or a pointer for
+                           submessages).
 ==================== ===== ================================================
 
 pb_field_t
@@ -61,7 +65,7 @@ Describes a single structure field with memory position in relation to others. T
 :tag:           Tag number of the field or 0 to terminate a list of fields.
 :type:          LTYPE and HTYPE of the field.
 :data_offset:   Offset of field data, relative to the end of the previous field.
-:size_offset:   Offset of *bool* flag for optional fields or *size_t* count for arrays, relative to field data.
+:size_offset:   Offset of *size_t* count for arrays, relative to field data.
 :data_size:     Size of a single data entry, in bytes. For PB_LTYPE_BYTES, the size of the byte array inside the containing structure. For PB_HTYPE_CALLBACK, size of the C data type if known.
 :array_size:    Maximum number of entries in an array, if it is an array type.
 :ptr:           Pointer to default value for optional fields, or to submessage description for PB_LTYPE_SUBMESSAGE.
@@ -79,6 +83,16 @@ An byte array with a field for storing the length::
 
 In an actual array, the length of *bytes* may be different.
 
+pb_bytes_t
+----------
+A byte array with fields for storing the allocated and current lengths::
+
+    typedef struct {
+        size_t alloced;
+        size_t size;
+        uint8_t *bytes;
+    } pb_bytes_array_t;
+
 pb_callback_t
 -------------
 Part of a message structure, for fields with type PB_HTYPE_CALLBACK::
@@ -88,6 +102,7 @@ Part of a message structure, for fields with type PB_HTYPE_CALLBACK::
         union {
             bool (*decode)(pb_istream_t *stream, const pb_field_t *field, void *arg);
             bool (*encode)(pb_ostream_t *stream, const pb_field_t *field, const void *arg);
+            bool (*encode_buffer)(pb_strstream_t *stream, const pb_field_t *field, const void *arg);
         } funcs;
         
         void *arg;
@@ -95,11 +110,11 @@ Part of a message structure, for fields with type PB_HTYPE_CALLBACK::
 
 The *arg* is passed to the callback when calling. It can be used to store any information that the callback might need.
 
-When calling `pb_encode`_, *funcs.encode* is used, and similarly when calling `pb_decode`_, *funcs.decode* is used. The function pointers are stored in the same memory location but are of incompatible types. You can set the function pointer to NULL to skip the field.
+When calling `pb_encode`_, *funcs.encode* is used, and similarly when calling `pb_encode_buffer`_, *funcs.encode_buffer* is used, and when calling `pb_decode`_, *funcs.decode* is used. The function pointers are stored in the same memory location but are of incompatible types. You can set the function pointer to NULL to skip the field.
 
 pb_wire_type_t
 --------------
-Protocol Buffers wire types. These are used with `pb_encode_tag`_. ::
+Protocol Buffers wire types. These are used with `pb_encode_tag`_ and `pb_encbuf_tag`_. ::
 
     typedef enum {
         PB_WT_VARINT = 0,
@@ -140,14 +155,14 @@ pb_encode
 ---------
 Encodes the contents of a structure as a protocol buffers message and writes it to output stream. ::
 
-    bool pb_encode(pb_ostream_t *stream, const pb_field_t fields[], const void *src_struct);
+    bool pb_encode(pb_ostream_t *stream, const pb_message_t *msg, const void *src_struct);
 
 :stream:        Output stream to write to.
-:fields:        A field description array, usually autogenerated.
+:msg:           A message descriptor, usually autogenerated.
 :src_struct:    Pointer to the data that will be serialized.
 :returns:       True on success, false on IO error, on detectable errors in field description, or if a field encoder returns false.
 
-Normally pb_encode simply walks through the fields description array and serializes each field in turn. However, submessages must be serialized twice: first to calculate their size and then to actually write them to output. This causes some constraints for callback fields, which must return the same data on every call.
+Normally pb_encode simply walks through the fields description array inside the message descriptor and serializes each field in turn. However, submessages must be serialized twice: first to calculate their size and then to actually write them to output. This causes some constraints for callback fields, which must return the same data on every call.
 
 pb_encode_varint
 ----------------
@@ -289,7 +304,7 @@ Field encoder for PB_LTYPE_SUBMESSAGE. Calls `pb_encode`_ to perform the actual
     bool pb_enc_submessage(pb_ostream_t *stream, const pb_field_t *field, const void *src);
 
 :stream:        Output stream to write to.
-:field:         Field description structure. The *ptr* field must be a pointer to a field description array for the submessage.
+:field:         Field description structure. The *ptr* field must be a pointer to a valid *pb_message_t* descriptor for the submessage.
 :src:           Pointer to the structure where submessage data is.
 :returns:       True on success, false on IO errors, pb_encode errors or if submessage size changes between calls.
 
@@ -297,6 +312,107 @@ In Protocol Buffers format, the submessage size must be written before the subme
 
 If the submessage contains callback fields, the callback function might misbehave and write out a different amount of data on the second call. This situation is recognized and *false* is returned, but it is up to the caller to ensure that the receiver of the message does not interpret it as valid data.
 
+pb_encode_buffer.h
+==================
+
+An important note about this module is that data is written from the
+back of the buffer to the front.  That is, when you call
+*pb_buf_write()*, it will place the bytes (in the order you provide
+them) before the data currently in the buffer.
+
+pb_strstream_from_buffer
+------------------------
+Constructs a buffer descriptor. This is just a helper function, it doesn't do anything you couldn't do yourself in a callback function. ::
+
+    pb_strstream_t pb_strstream_from_buffer(uint8_t *buf, size_t bufsize);
+
+:buf:           Memory buffer to write into.
+:bufsize:       Maximum number of bytes to write.
+:returns:       The buffer descriptor.
+
+The descriptor only tracks the amount of space left; it does not count how many bytes have been written.
+
+pb_buf_write
+------------
+Prepends data to an in-memory buffer. Always use this function, instead of trying to manage the pointers inside the buffer descriptor. ::
+
+    bool pb_buf_write(pb_strstream_t *stream, const uint8_t *buf, size_t count);
+
+:stream:        Descriptor for buffer to write to.
+:buf:           Pointer to buffer with the data to be written.
+:count:         Number of bytes to write.
+:returns:       True on success, false if maximum length is exceeded.
+
+If there is not enough space, *stream* is not modified.
+
+pb_encode_buffer
+----------------
+Encodes the contents of a structure as a protocol buffers message and writes it to a buffer. ::
+
+    bool pb_encode_buffer(pb_strstream_t *stream, const pb_message_t *msg, const void *src_struct);
+
+:stream:        Descriptor for buffer to write to.
+:msg:           A message descriptor, usually autogenerated.
+:src_struct:    Pointer to the data that will be serialized.
+:returns:       True on success, false if the buffer is too small or if a field encoder returns false.
+
+pb_encbuf_varint
+----------------
+Encodes an unsigned integer in the varint_ format. ::
+
+    bool pb_encbuf_varint(pb_strstream_t *stream, uint64_t value);
+
+:stream:        Descriptor for buffer to write to. 1-10 bytes will be written.
+:value:         Value to encode.
+:returns:       True on success, false on IO error.
+
+.. _varint: http://code.google.com/apis/protocolbuffers/docs/encoding.html#varints
+
+pb_encbuf_tag
+-------------
+Finishes a field in the Protocol Buffers binary format: encodes the field number and the wire type of the data. ::
+
+    bool pb_encbuf_tag(pb_strstream_t *stream, pb_wire_type_t wiretype, int field_number);
+
+:stream:        Descriptor for buffer to write to. 1-5 bytes will be written.
+:wiretype:      PB_WT_VARINT, PB_WT_64BIT, PB_WT_STRING or PB_WT_32BIT
+:field_number:  Identifier for the field, defined in the .proto file.
+:returns:       True on success, false on IO error.
+
+pb_encbuf_tag_for_field
+-----------------------
+Same as `pb_encbuf_tag`_, except takes the parameters from a *pb_field_t* structure. ::
+
+    bool pb_encbuf_tag_for_field(pb_strstream_t *stream, const pb_field_t *field);
+
+:stream:        Descriptor for buffer to write to. 1-5 bytes will be written.
+:field:         Field description structure. Usually autogenerated.
+:returns:       True on success, false on IO error or unknown field type.
+
+This function only considers the LTYPE of the field. You can use it from your field callbacks, because the source generator writes correct LTYPE also for callback type fields.
+
+Wire type mapping is as follows:
+
+========================= ============
+LTYPEs                    Wire type
+========================= ============
+VARINT, SVARINT           PB_WT_VARINT
+FIXED64                   PB_WT_64BIT  
+STRING, BYTES, SUBMESSAGE PB_WT_STRING 
+FIXED32                   PB_WT_32BIT
+========================= ============
+
+pb_encbuf_string
+----------------
+Writes the length of a string as varint and then contents of the string. Used for writing fields with wire type PB_WT_STRING. ::
+
+    bool pb_encbuf_string(pb_strstream_t *stream, const uint8_t *buffer, size_t size);
+
+:stream:        Descriptor for buffer to write to.
+:buffer:        Pointer to string data.
+:size:          Number of bytes in the string.
+:returns:       True on success, false on IO error.
+
 pb_decode.h
 ===========
 
@@ -355,10 +471,10 @@ pb_decode
 ---------
 Read and decode all fields of a structure. Reads until EOF on input stream. ::
 
-    bool pb_decode(pb_istream_t *stream, const pb_field_t fields[], void *dest_struct);
+    bool pb_decode(pb_istream_t *stream, const pb_message_t *msg, void *dest_struct);
 
 :stream:        Input stream to read from.
-:fields:        A field description array. Usually autogenerated.
+:msg:           A message descriptor. Usually autogenerated.
 :dest_struct:   Pointer to structure where data will be stored.
 :returns:       True on success, false on IO error, on detectable errors in field description, if a field encoder returns false or if a required field is missing.
 
@@ -366,9 +482,23 @@ In Protocol Buffers binary format, EOF is only allowed between fields. If it hap
 
 In addition to EOF, the pb_decode implementation supports terminating a message with a 0 byte. This is compatible with the official Protocol Buffers because 0 is never a valid field tag.
 
-For optional fields, this function applies the default value and sets *has_<field>* to false if the field is not present.
+For optional fields, this function applies the default value and clears the corresponding bit in *has_fields* if the field is not present.
 
-Because of memory concerns, the detection of missing required fields is not perfect if the structure contains more than 32 fields.
+pb_clean
+--------
+Release and clear all the unused pointers of a structure. ::
+
+    bool pb_clean(const pb_message_t *msg, void *dest_struct);
+
+:msg:           A message descriptor. Usually autogenerated.
+:dest_struct:   Pointer to structure to be cleaned.
+:returns:       True on success, false if one of the unused message fields has a pointer type that this function cannot handle.
+
+For each string or submessage with pointer type, this function calls *free()* on the pointer and sets the pointer to NULL.
+
+For bytes fields with pointer type, this function calls *free()* on the allocated data, nulls that pointer and zeros the size fields in the generated pb_bytes_t structure.
+
+For repeated fields, it applies the corresponding operation above to each unused element of the generated array.
 
 .. sidebar:: Field decoders
     
@@ -459,5 +589,5 @@ Field decoder for PB_LTYPE_SUBMESSAGE. Calls `pb_decode`_ to perform the actual
 :dest:          Pointer to the destination structure.
 :returns:       True on success, false on IO error or if `pb_decode`_ fails.
 
-The *field->ptr* should be a pointer to *pb_field_t* array describing the submessage.
+The *field->ptr* should be a pointer to *pb_message_t* describing the submessage.
 

example/client.c

@@ -27,7 +27,7 @@ bool printfile_callback(pb_istream_t *stream, const pb_field_t *field, void *arg
 {
     FileInfo fileinfo;
     
-    if (!pb_decode(stream, FileInfo_fields, &fileinfo))
+    if (!pb_decode(stream, FileInfo_msg, &fileinfo))
         return false;
     
     printf("%-10lld %s\n", fileinfo.inode, fileinfo.name);
@@ -45,11 +45,11 @@ bool listdir(int fd, char *path)
     
     if (path == NULL)
     {
-        request.has_path = false;
+        ListFilesRequest_clear(request, path);
     }
     else
     {
-        request.has_path = true;
+        ListFilesRequest_set(request, path);
         if (strlen(path) + 1 > sizeof(request.path))
         {
             fprintf(stderr, "Too long path.\n");
@@ -59,7 +59,7 @@ bool listdir(int fd, char *path)
         strcpy(request.path, path);
     }
     
-    if (!pb_encode(&output, ListFilesRequest_fields, &request))
+    if (!pb_encode(&output, ListFilesRequest_msg, &request))
     {
         fprintf(stderr, "Encoding failed.\n");
         return false;
@@ -70,7 +70,7 @@ bool listdir(int fd, char *path)
     
     response.file.funcs.decode = &printfile_callback;
     
-    if (!pb_decode(&input, ListFilesResponse_fields, &response))
+    if (!pb_decode(&input, ListFilesResponse_msg, &response))
     {
         fprintf(stderr, "Decoding failed.\n");
         return false;

example/server.c

@@ -53,7 +53,7 @@ void handle_connection(int connfd)
     pb_ostream_t output = pb_ostream_from_socket(connfd);
     DIR *directory;
     
-    if (!pb_decode(&input, ListFilesRequest_fields, &request))
+    if (!pb_decode(&input, ListFilesRequest_msg, &request))
     {
         printf("Decoding failed.\n");
         return;
@@ -67,18 +67,18 @@ void handle_connection(int connfd)
     {
         perror("opendir");
         
-        response.has_path_error = true;
+        ListFilesResponse_set(response, path_error);
         response.path_error = true;
         response.file.funcs.encode = NULL;
     }
     else
     {
-        response.has_path_error = false;
+        ListFilesResponse_clear(response, path_error);
         response.file.funcs.encode = &listdir_callback;
         response.file.arg = directory;
     }
     
-    if (!pb_encode(&output, ListFilesResponse_fields, &response))
+    if (!pb_encode(&output, ListFilesResponse_msg, &response))
     {
         printf("Encoding failed.\n");
     }

generator/nanopb.proto

@@ -1,6 +1,7 @@
 // Custom options for defining:
 // - Maximum size of string/bytes
 // - Maximum number of elements in array
+// - Pointer or in-line representation of non-scalar fields
 //
 // These are used by nanopb to generate statically allocable structures
 // for memory-limited environments.
@@ -10,6 +11,7 @@ import "google/protobuf/descriptor.proto";
 message NanoPBOptions {
   optional int32 max_size = 1;
   optional int32 max_count = 2;
+  optional bool pointer = 3;
 }
 
 // Protocol Buffers extension number registry

generator/nanopb_generator.py

@@ -3,6 +3,7 @@
 import google.protobuf.descriptor_pb2 as descriptor
 import nanopb_pb2
 import os.path
+from optparse import OptionParser
 
 # Values are tuple (c type, pb ltype)
 FieldD = descriptor.FieldDescriptorProto
@@ -21,6 +22,12 @@ datatypes = {
     FieldD.TYPE_UINT32: ('uint32_t', 'PB_LTYPE_VARINT'),
     FieldD.TYPE_UINT64: ('uint64_t', 'PB_LTYPE_VARINT')
 }
+pointable_types = frozenset([
+    FieldD.TYPE_STRING,
+    FieldD.TYPE_BYTES,
+    FieldD.TYPE_MESSAGE
+])
+options = None
 
 class Names:
     '''Keeps a set of nested names and formats them to C identifier.
@@ -70,6 +77,8 @@ class Field:
         self.max_size = None
         self.max_count = None
         self.array_decl = ""
+        self.is_pointer = options.pointer
+        self.is_array = False
         
         # Parse nanopb-specific field options
         if desc.options.HasExtension(nanopb_pb2.nanopb):
@@ -78,6 +87,11 @@ class Field:
                 self.max_size = ext.max_size
             if ext.HasField("max_count"):
                 self.max_count = ext.max_count
+            if ext.HasField("pointer"):
+                self.is_pointer = ext.pointer
+                if desc.type not in pointable_types:
+                    raise NotImplementedError('Cannot make %s.%s a pointer'
+                                              % (struct_name, desc.name))
         
         if desc.HasField('default_value'):
             self.default = desc.default_value
@@ -96,8 +110,11 @@ class Field:
             else:
                 self.htype = 'PB_HTYPE_ARRAY'
                 self.array_decl = '[%d]' % self.max_count
+                self.is_array = True
         else:
             raise NotImplementedError(desc.label)
+        if self.is_pointer:
+            self.htype += ' | PB_HTYPE_POINTER'
         
         # Decide LTYPE and CTYPE
         # LTYPE is the low-order nibble of nanopb field description,
@@ -112,14 +129,18 @@ class Field:
                 self.default = self.ctype + self.default
         elif desc.type == FieldD.TYPE_STRING:
             self.ltype = 'PB_LTYPE_STRING'
-            if self.max_size is None:
+            if self.is_pointer:
+                self.ctype = 'char'
+            elif self.max_size is None:
                 is_callback = True
             else:
                 self.ctype = 'char'
                 self.array_decl += '[%d]' % self.max_size
         elif desc.type == FieldD.TYPE_BYTES:
             self.ltype = 'PB_LTYPE_BYTES'
-            if self.max_size is None:
+            if self.is_pointer:
+                self.ctype = 'pb_bytes_t'
+            elif self.max_size is None:
                 is_callback = True
             else:
                 self.ctype = self.struct_name + self.name + 't'
@@ -138,13 +159,14 @@ class Field:
         return cmp(self.tag, other.tag)
     
     def __str__(self):
-        if self.htype == 'PB_HTYPE_OPTIONAL':
-            result = '    bool has_' + self.name + ';\n'
-        elif self.htype == 'PB_HTYPE_ARRAY':
+        if self.is_array:
             result = '    size_t ' + self.name + '_count;\n'
         else:
             result = ''
-        result += '    %s %s%s;' % (self.ctype, self.name, self.array_decl)
+        if self.is_pointer and self.ctype != 'pb_bytes_t':
+            result += '    %s *%s%s;' % (self.ctype, self.name, self.array_decl)
+        else:
+            result += '    %s %s%s;' % (self.ctype, self.name, self.array_decl)
         return result
     
     def types(self):
@@ -207,19 +229,16 @@ class Field:
         else:
             result += '    pb_delta_end(%s, %s, %s),' % (self.struct_name, self.name, prev_field_name)
         
-        if self.htype == 'PB_HTYPE_OPTIONAL':
-            result += '\n    pb_delta(%s, has_%s, %s),' % (self.struct_name, self.name, self.name)
-        elif self.htype == 'PB_HTYPE_ARRAY':
+        if self.is_array:
             result += '\n    pb_delta(%s, %s_count, %s),' % (self.struct_name, self.name, self.name)
         else:
             result += ' 0,'
         
-        
-        if self.htype == 'PB_HTYPE_ARRAY':
+        if self.is_array:
             result += '\n    pb_membersize(%s, %s[0]),' % (self.struct_name, self.name)
             result += ('\n    pb_membersize(%s, %s) / pb_membersize(%s, %s[0]),'
                        % (self.struct_name, self.name, self.struct_name, self.name))
-        elif self.htype != 'PB_HTYPE_CALLBACK' and self.ltype == 'PB_LTYPE_BYTES':
+        elif self.htype != 'PB_HTYPE_CALLBACK' and not self.is_pointer and self.ltype == 'PB_LTYPE_BYTES':
             result += '\n    pb_membersize(%s, bytes),' % self.ctype
             result += ' 0,'
         else:
@@ -227,7 +246,7 @@ class Field:
             result += ' 0,'
         
         if self.ltype == 'PB_LTYPE_SUBMESSAGE':
-            result += '\n    &%s_fields}' % self.submsgname
+            result += '\n    %s_msg}' % self.submsgname
         elif self.default is None or self.htype == 'PB_HTYPE_CALLBACK':
             result += ' 0}'
         else:
@@ -244,12 +263,14 @@ class Message:
 
     def get_dependencies(self):
         '''Get list of type names that this structure refers to.'''
-        return [str(field.ctype) for field in self.fields]
+        return [str(field.ctype) for field in self.fields if not
+                (field.is_pointer or field.htype == 'PB_HTYPE_CALLBACK')]
     
     def __str__(self):
-        result = 'typedef struct {\n'
+        result = 'struct %s {\n' % self.name
+        result += '    uint8_t has_fields[%d];\n' % ((len(self.fields) + 7) / 8)
         result += '\n'.join([str(f) for f in self.ordered_fields])
-        result += '\n} %s;' % self.name
+        result += '\n};'
         return result
     
     def types(self):
@@ -260,6 +281,9 @@ class Message:
                 result += types + '\n'
         return result
     
+    def typedef(self):
+        return 'typedef struct %s %s;' % (self.name, self.name)
+
     def default_decl(self, declaration_only = False):
         result = ""
         for field in self.fields:
@@ -268,20 +292,36 @@ class Message:
                 result += default + '\n'
         return result
 
-    def fields_declaration(self):
-        result = 'extern const pb_field_t %s_fields[%d];' % (self.name, len(self.fields) + 1)
+    def message_declaration(self):
+        result = 'typedef PB_MSG_STRUCT(%d) %s_msg_t;\n' % (len(self.fields), self.name)
+        result += 'extern const %s_msg_t %s_real_msg;\n' % (self.name, self.name)
+        result += '#define %s_msg ((const pb_message_t*)&%s_real_msg)\n' % (self.name, self.name)
         return result
 
-    def fields_definition(self):
-        result = 'const pb_field_t %s_fields[%d] = {\n' % (self.name, len(self.fields) + 1)
+    def message_definition(self):
+        result = 'const %s_msg_t %s_real_msg = {\n' % (self.name, self.name)
+        result += '    %d, sizeof(%s),\n' % (len(self.fields), self.name)
         
+        result += '    {\n\n'
         prev = None
         for field in self.ordered_fields:
             result += field.pb_field_t(prev)
             result += ',\n\n'
             prev = field.name
         
-        result += '    PB_LAST_FIELD\n};'
+        result += '    }\n};'
+        return result
+
+    def field_numbers(self):
+        result = '#define %s_has(STRUCT, FIELD) PB_HAS_FIELD(STRUCT, %s, FIELD)\n' % (self.name, self.name)
+        result += '#define %s_set(STRUCT, FIELD) PB_SET_FIELD(STRUCT, %s, FIELD)\n' % (self.name, self.name)
+        result += '#define %s_clear(STRUCT, FIELD) PB_CLEAR_FIELD(STRUCT, %s, FIELD)\n' % (self.name, self.name)
+
+        i = 0
+        for field in self.ordered_fields:
+            result += '#define %s_%s_index %d\n' % (self.name, field.name, i)
+            result += '#define %s_%s_tag %d\n' % (self.name, field.name, field.tag)
+            i += 1
         return result
 
 def iterate_messages(desc, names = Names()):
@@ -366,6 +406,11 @@ def generate_header(headername, enums, messages):
     for enum in enums:
         yield str(enum) + '\n\n'
     
+    yield '/* Struct typedefs */\n'
+    for msg in messages:
+        yield msg.typedef() + '\n'
+    yield '\n'
+    
     yield '/* Struct definitions */\n'
     for msg in sort_dependencies(messages):
         yield msg.types()
@@ -378,8 +423,13 @@ def generate_header(headername, enums, messages):
     
     yield '/* Struct field encoding specification for nanopb */\n'
     for msg in messages:
-        yield msg.fields_declaration() + '\n'
+        yield msg.message_declaration() + '\n'
+    yield '\n'
     
+    yield '/* Field indexes and tags */\n'
+    for msg in messages:
+        yield msg.field_numbers()
+
     yield '\n#endif\n'
 
 def generate_source(headername, enums, messages):
@@ -394,24 +444,31 @@ def generate_source(headername, enums, messages):
     yield '\n\n'
     
     for msg in messages:
-        yield msg.fields_definition() + '\n\n'
+        yield msg.message_definition() + '\n\n'
 
 if __name__ == '__main__':
     import sys
     import os.path
     
-    if len(sys.argv) != 2:
-        print "Usage: " + sys.argv[0] + " file.pb"
-        print "where file.pb has been compiled from .proto by:"
-        print "protoc -ofile.pb file.proto"
-        print "Output fill be written to file.pb.h and file.pb.c"
-        sys.exit(1)
-    
-    data = open(sys.argv[1]).read()
+    parser = OptionParser(usage="Usage: %prog [options] file.pb",
+                          epilog=
+"""file.pb should be compiled from file.proto by:
+protoc -ofile.pb file.proto
+""")
+    parser.add_option("-p", "--pointer", dest="pointer",
+                      action="store_true", default=False,
+                      help="generate pointers for non-scalar fields")
+
+    (options, args) = parser.parse_args()
+
+    if len(args) != 1:
+        parser.error("Please provide exactly one file.pb argument")
+
+    data = open(args[0]).read()
     fdesc = descriptor.FileDescriptorSet.FromString(data)
     enums, messages = parse_file(fdesc.file[0])
     
-    noext = os.path.splitext(sys.argv[1])[0]
+    noext = os.path.splitext(args[0])[0]
     headername = noext + '.pb.h'
     sourcename = noext + '.pb.c'
     headerbasename = os.path.basename(headername)

generator/nanopb_pb2.py

@@ -7,10 +7,12 @@ from google.protobuf import descriptor_pb2
 # @@protoc_insertion_point(imports)
 
 
+import google.protobuf.descriptor_pb2
+
 DESCRIPTOR = descriptor.FileDescriptor(
   name='nanopb.proto',
   package='',
-  serialized_pb='\n\x0cnanopb.proto\x1a google/protobuf/descriptor.proto\"4\n\rNanoPBOptions\x12\x10\n\x08max_size\x18\x01 \x01(\x05\x12\x11\n\tmax_count\x18\x02 \x01(\x05:>\n\x06nanopb\x12\x1d.google.protobuf.FieldOptions\x18\xf2\x07 \x01(\x0b\x32\x0e.NanoPBOptions')
+  serialized_pb='\n\x0cnanopb.proto\x1a google/protobuf/descriptor.proto\"E\n\rNanoPBOptions\x12\x10\n\x08max_size\x18\x01 \x01(\x05\x12\x11\n\tmax_count\x18\x02 \x01(\x05\x12\x0f\n\x07pointer\x18\x03 \x01(\x08:>\n\x06nanopb\x12\x1d.google.protobuf.FieldOptions\x18\xf2\x07 \x01(\x0b\x32\x0e.NanoPBOptions')
 
 
 NANOPB_FIELD_NUMBER = 1010
@@ -44,6 +46,13 @@ _NANOPBOPTIONS = descriptor.Descriptor(
       message_type=None, enum_type=None, containing_type=None,
       is_extension=False, extension_scope=None,
       options=None),
+    descriptor.FieldDescriptor(
+      name='pointer', full_name='NanoPBOptions.pointer', index=2,
+      number=3, type=8, cpp_type=7, label=1,
+      has_default_value=False, default_value=False,
+      message_type=None, enum_type=None, containing_type=None,
+      is_extension=False, extension_scope=None,
+      options=None),
   ],
   extensions=[
   ],
@@ -54,11 +63,10 @@ _NANOPBOPTIONS = descriptor.Descriptor(
   is_extendable=False,
   extension_ranges=[],
   serialized_start=50,
-  serialized_end=102,
+  serialized_end=119,
 )
 
-import google.protobuf.descriptor_pb2
-
+DESCRIPTOR.message_types_by_name['NanoPBOptions'] = _NANOPBOPTIONS
 
 class NanoPBOptions(message.Message):
   __metaclass__ = reflection.GeneratedProtocolMessageType

pb.h

@@ -27,58 +27,65 @@
  * SINT* is different, though, because it is zig-zag coded.
  */
 
-typedef enum {
-    /************************
-     * Field contents types *
-     ************************/
-    
-    /* Numeric types */
-    PB_LTYPE_VARINT = 0x00, /* int32, uint32, int64, uint64, bool, enum */
-    PB_LTYPE_SVARINT = 0x01, /* sint32, sint64 */
-    PB_LTYPE_FIXED32 = 0x02, /* fixed32, sfixed32, float */
-    PB_LTYPE_FIXED64 = 0x03, /* fixed64, sfixed64, double */
-    
-    /* Marker for last packable field type. */
-    PB_LTYPE_LAST_PACKABLE = 0x03,
-    
-    /* Byte array with pre-allocated buffer.
-     * data_size is the length of the allocated PB_BYTES_ARRAY structure. */
-    PB_LTYPE_BYTES = 0x04,
-    
-    /* String with pre-allocated buffer.
-     * data_size is the maximum length. */
-    PB_LTYPE_STRING = 0x05,
-    
-    /* Submessage
-     * submsg_fields is pointer to field descriptions */
-    PB_LTYPE_SUBMESSAGE = 0x06,
-    
-    /* Number of declared LTYPES */
-    PB_LTYPES_COUNT = 7,
-    
-    /******************
-     * Modifier flags *
-     ******************/
-    
-    /* Just the basic, write data at data_offset */
-    PB_HTYPE_REQUIRED = 0x00,
-    
-    /* Write true at size_offset */
-    PB_HTYPE_OPTIONAL = 0x10,
-    
-    /* Read to pre-allocated array
-     * Maximum number of entries is array_size,
-     * actual number is stored at size_offset */
-    PB_HTYPE_ARRAY = 0x20,
-    
-    /* Works for all required/optional/repeated fields.
-     * data_offset points to pb_callback_t structure.
-     * LTYPE should be 0 (it is ignored, but sometimes
-     * used to speculatively index an array). */
-    PB_HTYPE_CALLBACK = 0x30
-} pb_packed pb_type_t;
+/************************
+ * Field contents types *
+ ************************/
 
-#define PB_HTYPE(x) ((x) & 0xF0)
+typedef uint8_t pb_type_t;
+
+/* Numeric types */
+#define PB_LTYPE_VARINT 0x00 /* int32, uint32, int64, uint64, bool, enum */
+#define PB_LTYPE_SVARINT 0x01 /* sint32, sint64 */
+#define PB_LTYPE_FIXED32 0x02 /* fixed32, sfixed32, float */
+#define PB_LTYPE_FIXED64 0x03 /* fixed64, sfixed64, double */
+
+/* Marker for last packable field type. */
+#define PB_LTYPE_LAST_PACKABLE 0x03
+
+/* Byte array with pre-allocated buffer.
+ * data_size is the length of the allocated PB_BYTES_ARRAY structure. */
+#define PB_LTYPE_BYTES 0x04
+
+/* String with pre-allocated buffer.
+ * data_size is the maximum length. */
+#define PB_LTYPE_STRING 0x05
+
+/* Submessage
+ * submsg_fields is pointer to field descriptions */
+#define PB_LTYPE_SUBMESSAGE 0x06
+
+/* Number of declared LTYPES */
+#define PB_LTYPES_COUNT 7
+
+/******************
+ * Modifier flags *
+ ******************/
+
+/* Just the basic, write data at data_offset */
+#define PB_HTYPE_REQUIRED 0x00
+
+/* Write true at size_offset */
+#define PB_HTYPE_OPTIONAL 0x10
+
+/* Read to pre-allocated array
+ * Maximum number of entries is array_size,
+ * actual number is stored at size_offset */
+#define PB_HTYPE_ARRAY 0x20
+
+/* Works for all required/optional/repeated fields.
+ * data_offset points to pb_callback_t structure.
+ * LTYPE should be 0 (it is ignored, but sometimes
+ * used to speculatively index an array). */
+#define PB_HTYPE_CALLBACK 0x30
+
+/* Indicates that a string, bytes or non-repeated submessage is
+ * represented using a pointer (char* for string, pb_byptes_array_t
+ * for bytes).
+ */
+#define PB_HTYPE_POINTER 0x80
+
+#define PB_POINTER(x) ((x) & PB_HTYPE_POINTER)
+#define PB_HTYPE(x) ((x) & 0x70)
 #define PB_LTYPE(x) ((x) & 0x0F)
 
 /* This structure is used in auto-generated constants
@@ -97,7 +104,7 @@ struct _pb_field_t {
     uint8_t data_size; /* Data size in bytes for a single item */
     uint8_t array_size; /* Maximum number of entries in array */
     
-    /* Field definitions for submessage
+    /* Pointer to message structure for submessage
      * OR default value for all other non-array, non-callback types
      * If null, then field will zeroed. */
     const void *ptr;
@@ -112,6 +119,44 @@ typedef struct {
     uint8_t bytes[1];
 } pb_bytes_array_t;
 
+/* This structure is used for dynamically allocated 'bytes' arrays.
+ */
+typedef struct {
+    size_t size;
+    size_t alloced;
+    uint8_t *bytes;
+} pb_bytes_t;
+
+/* This macro defines the type of a structure for a message with N
+ * fields.
+ */
+#define PB_MSG_STRUCT(N) struct {               \
+        unsigned int field_count;               \
+        size_t size;                            \
+        pb_field_t fields[N];                   \
+    }
+
+/* This is the visible type for generated message structures.
+ * The actual number of fields at the end will vary by message type.
+ */
+typedef PB_MSG_STRUCT(1) pb_message_t;
+
+/* These macros are used to manipulate the has_fields array in
+ * generated messages.
+ */
+#define PB_FIELD_INDEX(TYPE, FIELD) TYPE ## _ ## FIELD ## _index
+#define PB_FIELD_BYTE(TYPE, FIELD) (PB_FIELD_INDEX(TYPE, FIELD) / 8)
+#define PB_FIELD_MASK(TYPE, FIELD) (1 << (PB_FIELD_INDEX(TYPE, FIELD) & 7))
+#define PB_HAS_FIELD(STRUCT, TYPE, FIELD) \
+    (((STRUCT).has_fields[PB_FIELD_BYTE(TYPE, FIELD)]   \
+      & PB_FIELD_MASK(TYPE, FIELD)) != 0)
+#define PB_SET_FIELD(STRUCT, TYPE, FIELD) \
+    ((STRUCT).has_fields[PB_FIELD_BYTE(TYPE, FIELD)]    \
+     |= PB_FIELD_MASK(TYPE, FIELD))
+#define PB_CLEAR_FIELD(STRUCT, TYPE, FIELD) \
+    ((STRUCT).has_fields[PB_FIELD_BYTE(TYPE, FIELD)]    \
+     &= ~PB_FIELD_MASK(TYPE, FIELD))
+
 /* This structure is used for giving the callback function.
  * It is stored in the message structure and filled in by the method that
  * calls pb_decode.
@@ -132,11 +177,13 @@ typedef struct {
  */
 typedef struct _pb_istream_t pb_istream_t;
 typedef struct _pb_ostream_t pb_ostream_t;
+typedef struct _pb_strstream_t pb_strstream_t;
 typedef struct _pb_callback_t pb_callback_t;
 struct _pb_callback_t {
     union {
         bool (*decode)(pb_istream_t *stream, const pb_field_t *field, void *arg);
         bool (*encode)(pb_ostream_t *stream, const pb_field_t *field, const void *arg);
+        bool (*encode_buffer)(pb_strstream_t *stream, const pb_field_t *field, const void *arg);
     } funcs;
     
     /* Free arg for use by callback */

pb_decode.c

@@ -14,6 +14,9 @@
 #include "pb.h"
 #include "pb_decode.h"
 #include <string.h>
+#ifdef MALLOC_HEADER
+# include MALLOC_HEADER
+#endif
 
 typedef bool (*pb_decoder_t)(pb_istream_t *stream, const pb_field_t *field, void *dest) checkreturn;
 
@@ -185,7 +188,7 @@ static bool checkreturn make_string_substream(pb_istream_t *stream, pb_istream_t
 
 /* Iterator for pb_field_t list */
 typedef struct {
-    const pb_field_t *start;
+    const pb_message_t *msg;
     const pb_field_t *current;
     int field_index;
     void *dest_struct;
@@ -193,9 +196,10 @@ typedef struct {
     void *pSize;
 } pb_field_iterator_t;
 
-static void pb_field_init(pb_field_iterator_t *iter, const pb_field_t *fields, void *dest_struct)
+static void pb_field_init(pb_field_iterator_t *iter, const pb_message_t *msg, void *dest_struct)
 {
-    iter->start = iter->current = fields;
+    iter->msg = msg;
+    iter->current = msg->fields;
     iter->field_index = 0;
     iter->pData = (char*)dest_struct + iter->current->data_offset;
     iter->pSize = (char*)iter->pData + iter->current->size_offset;
@@ -212,9 +216,9 @@ static bool pb_field_next(pb_field_iterator_t *iter)
     
     iter->current++;
     iter->field_index++;
-    if (iter->current->tag == 0)
+    if (iter->field_index >= iter->msg->field_count)
     {
-        iter->current = iter->start;
+        iter->current = &iter->msg->fields[0];
         iter->field_index = 0;
         iter->pData = iter->dest_struct;
         prev_size = 0;
@@ -247,13 +251,15 @@ static bool checkreturn decode_field(pb_istream_t *stream, pb_wire_type_t wire_t
 {
     pb_decoder_t func = PB_DECODERS[PB_LTYPE(iter->current->type)];
     
+#ifndef MALLOC_HEADER
+    if (PB_POINTER(iter->current->type))
+        return false;
+#endif
+    
     switch (PB_HTYPE(iter->current->type))
     {
         case PB_HTYPE_REQUIRED:
-            return func(stream, iter->current, iter->pData);
-            
         case PB_HTYPE_OPTIONAL:
-            *(bool*)iter->pSize = true;
             return func(stream, iter->current, iter->pData);
     
         case PB_HTYPE_ARRAY:
@@ -334,23 +340,22 @@ static bool checkreturn decode_field(pb_istream_t *stream, pb_wire_type_t wire_t
 }
 
 /* Initialize message fields to default values, recursively */
-static void pb_message_set_to_defaults(const pb_field_t fields[], void *dest_struct)
+static void pb_message_set_to_defaults(const pb_message_t *msg, void *dest_struct)
 {
     pb_field_iterator_t iter;
-    pb_field_init(&iter, fields, dest_struct);
+    pb_field_init(&iter, msg, dest_struct);
     
+    /* Initialize the has_fields array to zero. */
+    memset(dest_struct, 0, (msg->field_count + 7) / 8);
+
     /* Initialize size/has fields and apply default values */
     do
     {
-        if (iter.current->tag == 0)
+        if (iter.field_index >= msg->field_count)
             continue;
         
-        /* Initialize the size field for optional/repeated fields to 0. */
-        if (PB_HTYPE(iter.current->type) == PB_HTYPE_OPTIONAL)
-        {
-            *(bool*)iter.pSize = false;
-        }
-        else if (PB_HTYPE(iter.current->type) == PB_HTYPE_ARRAY)
+        /* Initialize the size field for repeated fields to 0. */
+        if (PB_HTYPE(iter.current->type) == PB_HTYPE_ARRAY)
         {
             *(size_t*)iter.pSize = 0;
             continue; /* Array is empty, no need to initialize contents */
@@ -363,6 +368,9 @@ static void pb_message_set_to_defaults(const pb_field_t fields[], void *dest_str
         }
         else if (PB_LTYPE(iter.current->type) == PB_LTYPE_SUBMESSAGE)
         {
+            if (PB_POINTER(iter.current->type) &&
+                (*(void**)iter.pData == NULL))
+                continue;
             pb_message_set_to_defaults(iter.current->ptr, iter.pData);
         }
         else if (iter.current->ptr != NULL)
@@ -380,15 +388,15 @@ static void pb_message_set_to_defaults(const pb_field_t fields[], void *dest_str
  * Decode all fields *
  *********************/
 
-bool checkreturn pb_decode(pb_istream_t *stream, const pb_field_t fields[], void *dest_struct)
+bool checkreturn pb_decode(pb_istream_t *stream, const pb_message_t *msg, void *dest_struct)
 {
-    uint32_t fields_seen = 0; /* Used to check for required fields */
     pb_field_iterator_t iter;
-    int i;
+    char *has_fields = dest_struct;
+    unsigned int i;
     
-    pb_message_set_to_defaults(fields, dest_struct);
+    pb_message_set_to_defaults(msg, dest_struct);
     
-    pb_field_init(&iter, fields, dest_struct);
+    pb_field_init(&iter, msg, dest_struct);
     
     while (stream->bytes_left)
     {
@@ -417,17 +425,17 @@ bool checkreturn pb_decode(pb_istream_t *stream, const pb_field_t fields[], void
             continue;
         }
         
-        fields_seen |= 1 << (iter.field_index & 31);
+        has_fields[iter.field_index/8] |= 1 << iter.field_index%8;
             
         if (!decode_field(stream, wire_type, &iter))
             return false;
     }
     
-    /* Check that all required fields (mod 31) were present. */
-    for (i = 0; fields[i].tag != 0; i++)
+    /* Check that all required fields were present. */
+    for (i = 0; i < msg->field_count; i++)
     {
-        if (PB_HTYPE(fields[i].type) == PB_HTYPE_REQUIRED &&
-            !(fields_seen & (1 << (i & 31))))
+        if (PB_HTYPE(msg->fields[i].type) == PB_HTYPE_REQUIRED &&
+            !(has_fields[i/8] & 1 << (i%8)))
         {
             return false;
         }
@@ -436,6 +444,85 @@ bool checkreturn pb_decode(pb_istream_t *stream, const pb_field_t fields[], void
     return true;
 }
 
+#ifdef MALLOC_HEADER
+
+/* Clean a single unused field (or unused array element). */
+static bool checkreturn pb_clean_pointer(const pb_field_t *field, void *data)
+{
+    switch (PB_LTYPE(field->type))
+    {
+    case PB_LTYPE_BYTES:
+    {
+        pb_bytes_t *bytes = (pb_bytes_t*)data;
+        bytes->size = 0;
+        bytes->alloced = 0;
+        free(bytes->bytes);
+        return true;
+    }
+    case PB_LTYPE_STRING:
+    case PB_LTYPE_SUBMESSAGE:
+    {
+        void **obj = (void**)data;
+        free(*obj);
+        *obj = NULL;
+        return true;
+    }
+    default:
+        return false;
+    }
+}
+
+/* Clean unused trailing elements in an array. */
+static bool checkreturn pb_clean_array(const pb_field_t *field, void *data, size_t count)
+{
+    unsigned int i;
+    for (i = count; i < field->array_size; i++)
+    {
+        if (!pb_clean_pointer(field, data))
+            return false;
+        data = (char*)data + field->data_size;
+    }
+    return true;
+}
+
+/* Clean unused fields in a message. */
+bool pb_clean(const pb_message_t *msg, void *dest_struct)
+{
+    const char *has_fields = dest_struct;
+    const void *pSize;
+    char *pData = (char*)dest_struct;
+    size_t prev_size = 0;
+    unsigned int i;
+    
+    for (i = 0; i < msg->field_count; i++)
+    {
+        const pb_field_t *field = &msg->fields[i];
+        bool status;
+        pData = pData + prev_size + field->data_offset;
+        pSize = pData + field->size_offset;
+        prev_size = field->data_size;
+        if (PB_HTYPE(field->type) == PB_HTYPE_ARRAY)
+            prev_size *= field->array_size;
+        
+        if (!PB_POINTER(field->type))
+            continue;
+        else if (PB_HTYPE(field->type) == PB_HTYPE_ARRAY)
+            status = pb_clean_array(field, pData, *(size_t*)pSize);
+        else if (!(has_fields[i/8] & 1 << (i%8)))
+            status = pb_clean_pointer(field, pData);
+        else /* pointer field is used */
+            continue;
+        
+        if (!status)
+            return false;
+    }
+    
+    return true;
+}
+
+#endif
+
+
 /* Field decoders */
 
 /* Copy destsize bytes from src so that values are casted properly.
@@ -512,6 +599,24 @@ bool checkreturn pb_dec_bytes(pb_istream_t *stream, const pb_field_t *field, voi
     if (x->size > field->data_size)
         return false;
     
+#ifdef MALLOC_HEADER
+    if (PB_POINTER(field->type))
+    {
+        pb_bytes_t *x2 = (pb_bytes_t*)dest;
+        
+        if (x2->alloced < x2->size)
+        {
+            void *new_bytes = realloc(x2->bytes, x2->size);
+            if (!new_bytes)
+                return false;
+            x2->alloced = x2->size;
+            x2->bytes = new_bytes;
+        }
+        
+        return pb_read(stream, x2->bytes, x2->size);
+    }
+#endif
+    
     return pb_read(stream, x->bytes, x->size);
 }
 
@@ -525,6 +630,18 @@ bool checkreturn pb_dec_string(pb_istream_t *stream, const pb_field_t *field, vo
     if (size > field->data_size - 1)
         return false;
     
+#ifdef MALLOC_HEADER
+    if (PB_POINTER(field->type))
+    {
+        uint8_t *string = (uint8_t*)realloc(*(uint8_t**)dest, size + 1);
+        if (!string)
+            return false;
+        
+        *(uint8_t**)dest = string;
+        dest = string;
+    }
+#endif
+    
     status = pb_read(stream, (uint8_t*)dest, size);
     *((uint8_t*)dest + size) = 0;
     return status;
@@ -534,6 +651,7 @@ bool checkreturn pb_dec_submessage(pb_istream_t *stream, const pb_field_t *field
 {
     bool status;
     pb_istream_t substream;
+    const pb_message_t *msg;
     
     if (!make_string_substream(stream, &substream))
         return false;
@@ -541,7 +659,26 @@ bool checkreturn pb_dec_submessage(pb_istream_t *stream, const pb_field_t *field
     if (field->ptr == NULL)
         return false;
     
-    status = pb_decode(&substream, (pb_field_t*)field->ptr, dest);
+    msg = (const pb_message_t*)field->ptr;
+    
+#ifdef MALLOC_HEADER
+    if (PB_POINTER(field->type))
+    {
+        if (*(void**)dest == NULL)
+        {
+            void *object = calloc(1, msg->size);
+            if (!object)
+                return false;
+            *(void**)dest = object;
+            pb_message_set_to_defaults(msg, object);
+            dest = object;
+        } else {
+            dest = *(void**)dest;
+        }
+    }
+#endif
+    
+    status = pb_decode(&substream, msg, dest);
     stream->state = substream.state;
     return status;
 }

pb_decode.h

@@ -40,9 +40,9 @@ bool pb_read(pb_istream_t *stream, uint8_t *buf, size_t count);
 
 /* Decode from stream to destination struct.
  * Returns true on success, false on any failure.
- * The actual struct pointed to by dest must match the description in fields.
+ * The actual struct pointed to by dest must match the description in msg.
  */
-bool pb_decode(pb_istream_t *stream, const pb_field_t fields[], void *dest_struct);
+bool pb_decode(pb_istream_t *stream, const pb_message_t *msg, void *dest_struct);
 
 /* --- Helper functions ---
  * You may want to use these from your caller or callbacks.
@@ -68,4 +68,10 @@ bool pb_dec_bytes(pb_istream_t *stream, const pb_field_t *field, void *dest);
 bool pb_dec_string(pb_istream_t *stream, const pb_field_t *field, void *dest);
 bool pb_dec_submessage(pb_istream_t *stream, const pb_field_t *field, void *dest);
 
+/* Release memory and clear pointers for any unused elements of
+ * dest_struct.  This function is only compiled when MALLOC_HEADER is
+ * defined for pb_decode.c.
+ */
+bool pb_clean(const pb_message_t *msg, void *dest_struct);
+
 #endif

pb_encode.c

@@ -140,15 +140,17 @@ static bool checkreturn encode_array(pb_ostream_t *stream, const pb_field_t *fie
     return true;
 }
 
-bool checkreturn pb_encode(pb_ostream_t *stream, const pb_field_t fields[], const void *src_struct)
+bool checkreturn pb_encode(pb_ostream_t *stream, const pb_message_t *msg, const void *src_struct)
 {
-    const pb_field_t *field = fields;
     const void *pData = src_struct;
     const void *pSize;
+    const char *has_fields = src_struct;
+    unsigned int i;
     size_t prev_size = 0;
     
-    while (field->tag != 0)
+    for (i = 0; i < msg->field_count; i++)
     {
+        const pb_field_t *field = &msg->fields[i];
         pb_encoder_t func = PB_ENCODERS[PB_LTYPE(field->type)];
         pData = (const char*)pData + prev_size + field->data_offset;
         pSize = (const char*)pData + field->size_offset;
@@ -159,6 +161,16 @@ bool checkreturn pb_encode(pb_ostream_t *stream, const pb_field_t fields[], cons
                 
         switch (PB_HTYPE(field->type))
         {
+            case PB_HTYPE_OPTIONAL:
+                if (!(has_fields[i/8] & (1 << i%8)))
+                    break;
+                if (PB_POINTER(field->type)
+                    && (PB_LTYPE(field->type) == PB_LTYPE_STRING
+                        || PB_LTYPE(field->type) == PB_LTYPE_SUBMESSAGE)
+                    && *(void**)pData == NULL)
+                    break;
+                /* else fall through to required case */
+            
             case PB_HTYPE_REQUIRED:
                 if (!pb_encode_tag_for_field(stream, field))
                     return false;
@@ -166,17 +178,6 @@ bool checkreturn pb_encode(pb_ostream_t *stream, const pb_field_t fields[], cons
                     return false;
                 break;
             
-            case PB_HTYPE_OPTIONAL:
-                if (*(bool*)pSize)
-                {
-                    if (!pb_encode_tag_for_field(stream, field))
-                        return false;
-                
-                    if (!func(stream, field, pData))
-                        return false;
-                }
-                break;
-            
             case PB_HTYPE_ARRAY:
                 if (!encode_array(stream, field, pData, *(size_t*)pSize, func))
                     return false;
@@ -193,8 +194,6 @@ bool checkreturn pb_encode(pb_ostream_t *stream, const pb_field_t fields[], cons
                 break;
             }
         }
-    
-        field++;
     }
     
     return true;
@@ -334,13 +333,22 @@ bool checkreturn pb_enc_fixed32(pb_ostream_t *stream, const pb_field_t *field, c
 
 bool checkreturn pb_enc_bytes(pb_ostream_t *stream, const pb_field_t *field, const void *src)
 {
-    pb_bytes_array_t *bytes = (pb_bytes_array_t*)src;
-    return pb_encode_string(stream, bytes->bytes, bytes->size);
+    if ((field != NULL) && PB_POINTER(field->type)) {
+        pb_bytes_t *bytes = (pb_bytes_t*)src;
+        return pb_encode_string(stream, bytes->bytes, bytes->size);
+    } else {
+        pb_bytes_array_t *bytes = (pb_bytes_array_t*)src;
+        return pb_encode_string(stream, bytes->bytes, bytes->size);
+    }
 }
 
 bool checkreturn pb_enc_string(pb_ostream_t *stream, const pb_field_t *field, const void *src)
 {
-    return pb_encode_string(stream, (uint8_t*)src, strlen((char*)src));
+    size_t len;
+    if ((field != NULL) && PB_POINTER(field->type))
+        src = *(char**)src;
+    len = src ? strlen((char*)src) : 0;
+    return pb_encode_string(stream, (uint8_t*)src, len);
 }
 
 bool checkreturn pb_enc_submessage(pb_ostream_t *stream, const pb_field_t *field, const void *src)
@@ -352,7 +360,12 @@ bool checkreturn pb_enc_submessage(pb_ostream_t *stream, const pb_field_t *field
     if (field->ptr == NULL)
         return false;
     
-    if (!pb_encode(&substream, (pb_field_t*)field->ptr, src))
+    if (PB_POINTER(field->type)) {
+        src = *(void**)src;
+        if (src == NULL)
+            return false;
+    }
+    if (!pb_encode(&substream, (const pb_message_t*)field->ptr, src))
         return false;
     
     size = substream.bytes_written;
@@ -373,7 +386,7 @@ bool checkreturn pb_enc_submessage(pb_ostream_t *stream, const pb_field_t *field
     substream.max_size = size;
     substream.bytes_written = 0;
     
-    status = pb_encode(&substream, (pb_field_t*)field->ptr, src);
+    status = pb_encode(&substream, (const pb_message_t*)field->ptr, src);
     
     stream->bytes_written += substream.bytes_written;
     stream->state = substream.state;

pb_encode.h

@@ -39,10 +39,10 @@ bool pb_write(pb_ostream_t *stream, const uint8_t *buf, size_t count);
 
 /* Encode struct to given output stream.
  * Returns true on success, false on any failure.
- * The actual struct pointed to by src_struct must match the description in fields.
+ * The actual struct pointed to by src_struct must match the description in msg.
  * All required fields in the struct are assumed to have been filled in.
  */
-bool pb_encode(pb_ostream_t *stream, const pb_field_t fields[], const void *src_struct);
+bool pb_encode(pb_ostream_t *stream, const pb_message_t *msg, const void *src_struct);
 
 /* --- Helper functions ---
  * You may want to use these from your caller or callbacks.

pb_encode_buffer.c

@@ -0,0 +1,362 @@
+/* pb_encode_buffer.c - encode a protobuf to an in-memory array
+ *
+ * 2011 Michael Poole <mdpoole@troilus.org>
+ * Part of nanopb, 2011 Petteri Aimonen <jpa@kapsi.fi>
+ */
+
+#include "pb_encode_buffer.h"
+#include <string.h>
+
+#ifdef __GNUC__
+/* Verify that we remember to check all return values for proper error propagation */
+#define checkreturn __attribute__((warn_unused_result))
+#else
+#define checkreturn
+#endif
+
+typedef bool (*pb_encoder_t)(pb_strstream_t *stream, const pb_field_t *field, const void *src) checkreturn;
+
+static bool pb_encb_varint(pb_strstream_t *stream, const pb_field_t *field, const void *src);
+static bool pb_encb_svarint(pb_strstream_t *stream, const pb_field_t *field, const void *src);
+static bool pb_encb_fixed32(pb_strstream_t *stream, const pb_field_t *field, const void *src);
+static bool pb_encb_fixed64(pb_strstream_t *stream, const pb_field_t *field, const void *src);
+static bool pb_encb_bytes(pb_strstream_t *stream, const pb_field_t *field, const void *src);
+static bool pb_encb_string(pb_strstream_t *stream, const pb_field_t *field, const void *src);
+static bool pb_encb_submessage(pb_strstream_t *stream, const pb_field_t *field, const void *src);
+
+/* --- Function pointers to field encoders ---
+ * Order in the array must match pb_action_t LTYPE numbering.
+ */
+static const pb_encoder_t PB_ENCODERS[PB_LTYPES_COUNT] = {
+    &pb_encb_varint,
+    &pb_encb_svarint,
+    &pb_encb_fixed32,
+    &pb_encb_fixed64,
+    
+    &pb_encb_bytes,
+    &pb_encb_string,
+    &pb_encb_submessage
+};
+
+/* pb_strstream_t implementation */
+
+pb_strstream_t pb_strstream_from_buffer(uint8_t *buf, size_t bufsize)
+{
+    pb_strstream_t stream;
+    stream.buffer = buf;
+    stream.last = buf + bufsize;
+    return stream;
+}
+
+bool pb_buf_write(pb_strstream_t *stream, const uint8_t *buf, size_t count)
+{
+    if (stream->buffer + count > stream->last)
+        return false;
+    stream->last -= count;
+    memcpy(stream->last, buf, count);
+    return true;
+}
+
+/* Main encoding stuff */
+
+/* This function is static for the same reason as the version in
+ * pb_encode.c.
+ */
+static bool checkreturn encode_array(pb_strstream_t *stream, const pb_field_t *field,
+                                     const void *pData, size_t count, pb_encoder_t func)
+{
+    const void *p;
+    uint8_t *start;
+    size_t size;
+    int i;
+    
+    if (count == 0)
+        return true;
+    
+    start = stream->last;
+    p = (const char*)pData + field->data_size * count;
+    
+    if (PB_LTYPE(field->type) <= PB_LTYPE_LAST_PACKABLE)
+    {
+        /* Write the data (in reverse order). */
+        for (i = 0; i < count; i++)
+        {
+            p = (const char*)p - field->data_size;
+            if (!func(stream, field, p))
+                return false;
+        }
+        
+        /* Write the size. */
+        size = start - stream->last;
+        if (!pb_encbuf_varint(stream, size))
+            return false;
+        if (!pb_encbuf_tag(stream, PB_WT_STRING, field->tag))
+            return false;
+    }
+    else
+    {
+        for (i = 0; i < count; i++)
+        {
+            p = (const char*)p - field->data_size;
+            if (!func(stream, field, p))
+                return false;
+            if (!pb_encbuf_tag_for_field(stream, field))
+                return false;
+        }
+    }
+    
+    return true;
+}
+
+bool checkreturn pb_encode_buffer(pb_strstream_t *stream, const pb_message_t *msg, const void *src_struct)
+{
+    const char *has_fields = src_struct;
+    const uint8_t *pData = src_struct;
+    const pb_field_t *field;
+    pb_encoder_t func;
+    size_t size;
+    unsigned int i;
+
+    /* msg->size includes trailing padding, so we must calculate the
+     * offset of the last field by counting forward from the start.
+     */
+    for (i = 0; i < msg->field_count; i++)
+    {
+        field = &msg->fields[i];
+        pData += field->data_offset;
+        size = field->data_size;
+        if (PB_HTYPE(field->type) == PB_HTYPE_ARRAY)
+            size *= field->array_size;
+        pData += size;
+    }
+
+    /* Iterate through the fields in reverse order.  Because we write
+     * from the end of the buffer, the result is in canonical order.
+     */
+    for (i = msg->field_count; i > 0; pData -= field->data_offset)
+    {
+        field = &msg->fields[--i];
+        func = PB_ENCODERS[PB_LTYPE(field->type)];
+        size = field->data_size;
+        if (PB_HTYPE(field->type) == PB_HTYPE_ARRAY)
+            size *= field->array_size;
+        pData -= size;
+        
+        switch (PB_HTYPE(field->type))
+        {
+        case PB_HTYPE_OPTIONAL:
+            if (!(has_fields[i/8] & (1 << i%8)))
+                break;
+            if (PB_POINTER(field->type)
+                && (PB_LTYPE(field->type) == PB_LTYPE_STRING
+                        || PB_LTYPE(field->type) == PB_LTYPE_SUBMESSAGE)
+                    && *(void**)pData == NULL)
+                    break;
+                /* else fall through to required case */
+
+        case PB_HTYPE_REQUIRED:
+            if (!func(stream, field, pData))
+                return false;
+            if (!pb_encbuf_tag_for_field(stream, field))
+                return false;
+            break;
+
+        case PB_HTYPE_ARRAY:
+            size = *(size_t*)(pData + field->size_offset);
+            if (!encode_array(stream, field, pData, size, func))
+                return false;
+            break;
+
+        case PB_HTYPE_CALLBACK:
+        {
+            pb_callback_t *callback = (pb_callback_t*)pData;
+            if (callback->funcs.encode_buffer != NULL
+                && !callback->funcs.encode_buffer(stream, field, callback->arg))
+                return false;
+            break;
+        }
+        }
+    }
+    
+    return true;
+}
+
+/* Helper functions */
+
+bool checkreturn pb_encbuf_varint(pb_strstream_t *stream, uint64_t value)
+{
+    if (stream->last == stream->buffer)
+    {
+        return false;
+    }
+    else if (value < 128)
+    {
+        *--stream->last = value;
+        return true;
+    }
+    else
+    {
+        uint8_t buffer[10];
+        int i = 0;
+
+        while (value)
+        {
+            buffer[i] = (value & 0x7F) | 0x80;
+            value >>= 7;
+            i++;
+        }
+        buffer[i-1] &= 0x7F;
+        return pb_buf_write(stream, buffer, i);
+    }
+}
+
+bool checkreturn pb_encbuf_tag(pb_strstream_t *stream, pb_wire_type_t wiretype, int field_number)
+{
+    int tag = wiretype | (field_number << 3);
+    return pb_encbuf_varint(stream, tag);
+}
+
+bool checkreturn pb_encbuf_tag_for_field(pb_strstream_t *stream, const pb_field_t *field)
+{
+    pb_wire_type_t wiretype;
+    switch (PB_LTYPE(field->type))
+    {
+        case PB_LTYPE_VARINT:
+        case PB_LTYPE_SVARINT:
+            wiretype = PB_WT_VARINT;
+            break;
+        
+        case PB_LTYPE_FIXED32:
+            wiretype = PB_WT_32BIT;
+            break;
+        
+        case PB_LTYPE_FIXED64:
+            wiretype = PB_WT_64BIT;
+            break;
+        
+        case PB_LTYPE_BYTES:
+        case PB_LTYPE_STRING:
+        case PB_LTYPE_SUBMESSAGE:
+            wiretype = PB_WT_STRING;
+            break;
+        
+        default:
+            return false;
+    }
+    
+    return pb_encbuf_tag(stream, wiretype, field->tag);
+}
+
+bool checkreturn pb_encbuf_string(pb_strstream_t *stream, const uint8_t *buffer, size_t size)
+{
+    if (!pb_buf_write(stream, buffer, size))
+        return false;
+    return pb_encbuf_varint(stream, size);
+}
+
+/* Field encoders */
+
+/* Copy srcsize bytes of integer from src so that values are casted properly.
+ * On little endian machine, copy to start of dest
+ * On big endian machine, copy to end of dest
+ * destsize must always be larger than srcsize
+ * 
+ * Note: This is the reverse of the endian_copy in pb_decode.c.
+ */
+static void endian_copy(uint64_t *dest, const void *src, size_t srcsize)
+{
+#ifdef __BIG_ENDIAN__
+    memcpy((char*)dest + sizeof(*dest) - srcsize, src, srcsize);
+#else
+    memcpy(dest, src, srcsize);
+#endif
+}
+
+static bool pb_encb_varint(pb_strstream_t *stream, const pb_field_t *field, const void *src)
+{
+    uint64_t value = 0;
+    endian_copy(&value, src, field->data_size);
+    return pb_encbuf_varint(stream, value);
+}
+
+static bool pb_encb_svarint(pb_strstream_t *stream, const pb_field_t *field, const void *src)
+{
+    uint64_t value = 0;
+    uint64_t zigzagged;
+    uint64_t signbitmask = (uint64_t)0x80 << (field->data_size * 8 - 8);
+    uint64_t xormask = ((uint64_t)-1) >> (64 - field->data_size * 8);
+    
+    endian_copy(&value, src, field->data_size);
+    if (value & signbitmask)
+        zigzagged = ((value ^ xormask) << 1) | 1;
+    else
+        zigzagged = value << 1;
+    
+    return pb_encbuf_varint(stream, zigzagged);
+}
+
+static bool pb_encb_fixed32(pb_strstream_t *stream, const pb_field_t *field, const void *src)
+{
+#ifdef __BIG_ENDIAN__
+    const uint8_t *bytes = (const uint8_t*)src;
+    uint8_t lebytes[4] = {bytes[3], bytes[2], bytes[1], bytes[0]};
+    src = lebytes;
+#endif
+    return pb_buf_write(stream, src, 4);
+}
+
+static bool pb_encb_fixed64(pb_strstream_t *stream, const pb_field_t *field, const void *src)
+{
+#ifdef __BIG_ENDIAN__
+    const uint8_t *bytes = (const uint8_t*)src;
+    uint8_t lebytes[8] = {bytes[7], bytes[6], bytes[5], bytes[4],
+                          bytes[3], bytes[2], bytes[1], bytes[0]};
+    src = lebytes;
+#endif
+    return pb_buf_write(stream, src, 8);
+}
+
+static bool pb_encb_bytes(pb_strstream_t *stream, const pb_field_t *field, const void *src)
+{
+    if ((field != NULL) && PB_POINTER(field->type)) {
+        pb_bytes_t *bytes = (pb_bytes_t*)src;
+        return pb_encbuf_string(stream, bytes->bytes, bytes->size);
+    } else {
+        pb_bytes_array_t *bytes = (pb_bytes_array_t*)src;
+        return pb_encbuf_string(stream, bytes->bytes, bytes->size);
+    }
+}
+
+static bool pb_encb_string(pb_strstream_t *stream, const pb_field_t *field, const void *src)
+{
+    size_t len;
+    if ((field != NULL) && PB_POINTER(field->type))
+        src = *(char**)src;
+    len = src ? strlen((char*)src) : 0;
+    return pb_encbuf_string(stream, (uint8_t*)src, len);
+}
+
+static bool pb_encb_submessage(pb_strstream_t *stream, const pb_field_t *field, const void *src)
+{
+    uint8_t *start;
+    size_t size;
+    
+    if (field->ptr == NULL)
+        return false;
+    
+    if (PB_POINTER(field->type)) {
+        src = *(void**)src;
+        if (src == NULL)
+            return false;
+    }
+    
+    start = stream->last;
+    if (!pb_encode_buffer(stream, (const pb_message_t*)field->ptr, src))
+        return false;
+    
+    size = start - stream->last;
+    if (!pb_encbuf_varint(stream, size))
+        return false;
+    
+    return true;
+}

pb_encode_buffer.h

@@ -0,0 +1,43 @@
+#ifndef _PB_ENCODE_BUFFER_H_
+#define _PB_ENCODE_BUFFER_H_
+
+/* pb_encode_buffer.h: Functions to encode protocol buffers to an
+ * in-memory array of bytes. Depends on pb_encode_buffer.c. The main
+ * function is pb_encode_buffer.  You also need an output buffer,
+ * structures and their field descriptions (just like with pb_decode
+ * or pb_encode).
+ */
+
+#include <stdbool.h>
+#include "pb.h"
+
+/* Output stream for an in-memory buffer.
+ */
+struct _pb_strstream_t
+{
+    uint8_t *buffer;
+    uint8_t *last;
+};
+
+pb_strstream_t pb_strstream_from_buffer(uint8_t *buf, size_t bufsize);
+bool pb_buf_write(pb_strstream_t *stream, const uint8_t *buf, size_t count);
+
+/* --- Helper functions ---
+ * You may want to use these from your caller or callbacks.
+ */
+
+bool pb_encbuf_varint(pb_strstream_t *stream, uint64_t value);
+bool pb_encbuf_tag(pb_strstream_t *stream, pb_wire_type_t wiretype, int field_number);
+/* Encode tag based on LTYPE and field number defined in the field structure. */
+bool pb_encbuf_tag_for_field(pb_strstream_t *stream, const pb_field_t *field);
+/* Write length as varint and then the contents of buffer. */
+bool pb_encbuf_string(pb_strstream_t *stream, const uint8_t *buffer, size_t size);
+
+/* Encode struct to given output stream.
+ * Returns true on success, false on any failure.
+ * The actual struct pointed to by src_struct must match the description in msg.
+ * All required fields in the struct are assumed to have been filled in.
+ */
+bool pb_encode_buffer(pb_strstream_t *stream, const pb_message_t *msg, const void *src_struct);
+
+#endif

tests/Makefile

@@ -1,12 +1,12 @@
 CFLAGS=-ansi -Wall -Werror -I .. -g -O0 --coverage
 LDFLAGS=--coverage
-DEPS=../pb_decode.h ../pb_encode.h ../pb.h person.pb.h callbacks.pb.h unittests.h unittestproto.pb.h
-TESTS=test_decode1 test_encode1 decode_unittests encode_unittests
+DEPS=../pb_decode.h ../pb_encode.h ../pb_encode_buffer.h ../pb.h person.pb.h callbacks.pb.h unittests.h unittestproto.pb.h
+TESTS=test_decode1 test_encode1 test_decode_callbacks test_encode_callbacks decode_unittests decode_ptr_unittests encode_unittests
 
 all: breakpoints $(TESTS) run_unittests
 
 clean:
-	rm -f $(TESTS) person.pb* *.o *.gcda *.gcno
+	rm -f $(TESTS) *.pb.c *.pb.h *.o *.gcda *.gcno
 
 %.o: %.c
 %.o: %.c $(DEPS)
@@ -14,8 +14,14 @@ clean:
 
 pb_encode.o: ../pb_encode.c $(DEPS)
 	$(CC) $(CFLAGS) -c -o $@ $<
+pb_encode_buffer.o: ../pb_encode_buffer.c $(DEPS)
+	$(CC) $(CFLAGS) -c -o $@ $<
 pb_decode.o: ../pb_decode.c $(DEPS)
 	$(CC) $(CFLAGS) -c -o $@ $<
+pb_ptr_decode.o: ../pb_decode.c $(DEPS)
+	$(CC) $(CFLAGS) -c -o $@ $<
+decode_ptr_unittests.o: decode_unittests.c $(DEPS)
+	$(CC) $(CFLAGS) -c -o $@ $<
 
 test_decode1: test_decode1.o pb_decode.o person.pb.o
 test_decode2: test_decode2.o pb_decode.o person.pb.o
@@ -24,7 +30,9 @@ test_encode2: test_encode2.o pb_encode.o person.pb.o
 test_decode_callbacks: test_decode_callbacks.o pb_decode.o callbacks.pb.o
 test_encode_callbacks: test_encode_callbacks.o pb_encode.o callbacks.pb.o
 decode_unittests: decode_unittests.o pb_decode.o unittestproto.pb.o
-encode_unittests: encode_unittests.o pb_encode.o unittestproto.pb.o
+pb_ptr_decode.o decode_ptr_unittests.o: CFLAGS += -DMALLOC_HEADER="<stdlib.h>"
+decode_ptr_unittests: decode_ptr_unittests.o pb_ptr_decode.o unittestproto.pb.o
+encode_unittests: encode_unittests.o pb_encode.o pb_encode_buffer.o unittestproto.pb.o
 
 %.pb: %.proto
 	protoc -I. -I../generator -I/usr/include -o$@ $<
@@ -37,12 +45,15 @@ breakpoints: ../*.c *.c
 
 coverage: run_unittests
 	gcov pb_encode.gcda
+	gcov pb_encode_buffer.gcda
 	gcov pb_decode.gcda
+	gcov pb_ptr_decode.gcda
 
-run_unittests: decode_unittests encode_unittests test_encode1 test_encode2 test_decode1 test_decode2 test_encode_callbacks test_decode_callbacks
+run_unittests: decode_unittests decode_ptr_unittests encode_unittests test_encode1 test_encode2 test_decode1 test_decode2 test_encode_callbacks test_decode_callbacks
 	rm -f *.gcda
 	
 	./decode_unittests > /dev/null
+	./decode_ptr_unittests > /dev/null
 	./encode_unittests > /dev/null
 	
 	[ "`./test_encode1 | ./test_decode1`" = \

tests/decode_unittests.c

@@ -193,13 +193,13 @@ int main()
         IntegerArray dest;
         
         COMMENT("Testing pb_decode with repeated int32 field")
-        TEST((s = S(""), pb_decode(&s, IntegerArray_fields, &dest) && dest.data_count == 0))
-        TEST((s = S("\x08\x01\x08\x02"), pb_decode(&s, IntegerArray_fields, &dest)
+        TEST((s = S(""), pb_decode(&s, IntegerArray_msg, &dest) && dest.data_count == 0))
+        TEST((s = S("\x08\x01\x08\x02"), pb_decode(&s, IntegerArray_msg, &dest)
              && dest.data_count == 2 && dest.data[0] == 1 && dest.data[1] == 2))
         s = S("\x08\x01\x08\x02\x08\x03\x08\x04\x08\x05\x08\x06\x08\x07\x08\x08\x08\x09\x08\x0A");
-        TEST(pb_decode(&s, IntegerArray_fields, &dest) && dest.data_count == 10 && dest.data[9] == 10)
+        TEST(pb_decode(&s, IntegerArray_msg, &dest) && dest.data_count == 10 && dest.data[9] == 10)
         s = S("\x08\x01\x08\x02\x08\x03\x08\x04\x08\x05\x08\x06\x08\x07\x08\x08\x08\x09\x08\x0A\x08\x0B");
-        TEST(!pb_decode(&s, IntegerArray_fields, &dest))
+        TEST(!pb_decode(&s, IntegerArray_msg, &dest))
     }
     
     {
@@ -207,17 +207,17 @@ int main()
         IntegerArray dest;
         
         COMMENT("Testing pb_decode with packed int32 field")
-        TEST((s = S("\x0A\x00"), pb_decode(&s, IntegerArray_fields, &dest)
+        TEST((s = S("\x0A\x00"), pb_decode(&s, IntegerArray_msg, &dest)
             && dest.data_count == 0))
-        TEST((s = S("\x0A\x01\x01"), pb_decode(&s, IntegerArray_fields, &dest)
+        TEST((s = S("\x0A\x01\x01"), pb_decode(&s, IntegerArray_msg, &dest)
             && dest.data_count == 1 && dest.data[0] == 1))
-        TEST((s = S("\x0A\x0A\x01\x02\x03\x04\x05\x06\x07\x08\x09\x0A"), pb_decode(&s, IntegerArray_fields, &dest)
+        TEST((s = S("\x0A\x0A\x01\x02\x03\x04\x05\x06\x07\x08\x09\x0A"), pb_decode(&s, IntegerArray_msg, &dest)
             && dest.data_count == 10 && dest.data[0] == 1 && dest.data[9] == 10))
-        TEST((s = S("\x0A\x0B\x01\x02\x03\x04\x05\x06\x07\x08\x09\x0A\x0B"), !pb_decode(&s, IntegerArray_fields, &dest)))
+        TEST((s = S("\x0A\x0B\x01\x02\x03\x04\x05\x06\x07\x08\x09\x0A\x0B"), !pb_decode(&s, IntegerArray_msg, &dest)))
         
         /* Test invalid wire data */
-        TEST((s = S("\x0A\xFF"), !pb_decode(&s, IntegerArray_fields, &dest)))
-        TEST((s = S("\x0A\x01"), !pb_decode(&s, IntegerArray_fields, &dest)))
+        TEST((s = S("\x0A\xFF"), !pb_decode(&s, IntegerArray_msg, &dest)))
+        TEST((s = S("\x0A\x01"), !pb_decode(&s, IntegerArray_msg, &dest)))
     }
     
     {
@@ -225,14 +225,14 @@ int main()
         IntegerArray dest;
         
         COMMENT("Testing pb_decode with unknown fields")
-        TEST((s = S("\x18\x0F\x08\x01"), pb_decode(&s, IntegerArray_fields, &dest)
+        TEST((s = S("\x18\x0F\x08\x01"), pb_decode(&s, IntegerArray_msg, &dest)
             && dest.data_count == 1 && dest.data[0] == 1))
-        TEST((s = S("\x19\x00\x00\x00\x00\x00\x00\x00\x00\x08\x01"), pb_decode(&s, IntegerArray_fields, &dest)
+        TEST((s = S("\x19\x00\x00\x00\x00\x00\x00\x00\x00\x08\x01"), pb_decode(&s, IntegerArray_msg, &dest)
             && dest.data_count == 1 && dest.data[0] == 1))
-        TEST((s = S("\x1A\x00\x08\x01"), pb_decode(&s, IntegerArray_fields, &dest)
+        TEST((s = S("\x1A\x00\x08\x01"), pb_decode(&s, IntegerArray_msg, &dest)
             && dest.data_count == 1 && dest.data[0] == 1))
-        TEST((s = S("\x1B\x08\x01"), !pb_decode(&s, IntegerArray_fields, &dest)))
-        TEST((s = S("\x1D\x00\x00\x00\x00\x08\x01"), pb_decode(&s, IntegerArray_fields, &dest)
+        TEST((s = S("\x1B\x08\x01"), !pb_decode(&s, IntegerArray_msg, &dest)))
+        TEST((s = S("\x1D\x00\x00\x00\x00\x08\x01"), pb_decode(&s, IntegerArray_msg, &dest)
             && dest.data_count == 1 && dest.data[0] == 1))
     }
     
@@ -246,25 +246,25 @@ int main()
         COMMENT("Testing pb_decode with callbacks")
         /* Single varint */
         ref.size = 1; ref.bytes[0] = 0x55;
-        TEST((s = S("\x08\x55"), pb_decode(&s, CallbackArray_fields, &dest)))
+        TEST((s = S("\x08\x55"), pb_decode(&s, CallbackArray_msg, &dest)))
         /* Packed varint */
         ref.size = 3; ref.bytes[0] = ref.bytes[1] = ref.bytes[2] = 0x55;
-        TEST((s = S("\x0A\x03\x55\x55\x55"), pb_decode(&s, CallbackArray_fields, &dest)))
+        TEST((s = S("\x0A\x03\x55\x55\x55"), pb_decode(&s, CallbackArray_msg, &dest)))
         /* Packed varint with loop */
         ref.size = 1; ref.bytes[0] = 0x55;
-        TEST((s = S("\x0A\x03\x55\x55\x55"), pb_decode(&s, CallbackArray_fields, &dest)))
+        TEST((s = S("\x0A\x03\x55\x55\x55"), pb_decode(&s, CallbackArray_msg, &dest)))
         /* Single fixed32 */
         ref.size = 4; ref.bytes[0] = ref.bytes[1] = ref.bytes[2] = ref.bytes[3] = 0xAA;
-        TEST((s = S("\x0D\xAA\xAA\xAA\xAA"), pb_decode(&s, CallbackArray_fields, &dest)))
+        TEST((s = S("\x0D\xAA\xAA\xAA\xAA"), pb_decode(&s, CallbackArray_msg, &dest)))
         /* Single fixed64 */
         ref.size = 8; memset(ref.bytes, 0xAA, 8);
-        TEST((s = S("\x09\xAA\xAA\xAA\xAA\xAA\xAA\xAA\xAA"), pb_decode(&s, CallbackArray_fields, &dest)))
+        TEST((s = S("\x09\xAA\xAA\xAA\xAA\xAA\xAA\xAA\xAA"), pb_decode(&s, CallbackArray_msg, &dest)))
         /* Unsupported field type */
-        TEST((s = S("\x0B\x00"), !pb_decode(&s, CallbackArray_fields, &dest)))
+        TEST((s = S("\x0B\x00"), !pb_decode(&s, CallbackArray_msg, &dest)))
         
         /* Just make sure that our test function works */
         ref.size = 1; ref.bytes[0] = 0x56;
-        TEST((s = S("\x08\x55"), !pb_decode(&s, CallbackArray_fields, &dest)))
+        TEST((s = S("\x08\x55"), !pb_decode(&s, CallbackArray_msg, &dest)))
     }
     
     {
@@ -272,11 +272,39 @@ int main()
         IntegerArray dest;
         
         COMMENT("Testing pb_decode message termination")
-        TEST((s = S(""), pb_decode(&s, IntegerArray_fields, &dest)))
-        TEST((s = S("\x00"), pb_decode(&s, IntegerArray_fields, &dest)))
-        TEST((s = S("\x08\x01"), pb_decode(&s, IntegerArray_fields, &dest)))
-        TEST((s = S("\x08\x01\x00"), pb_decode(&s, IntegerArray_fields, &dest)))
-        TEST((s = S("\x08"), !pb_decode(&s, IntegerArray_fields, &dest)))
+        TEST((s = S(""), pb_decode(&s, IntegerArray_msg, &dest)))
+        TEST((s = S("\x00"), pb_decode(&s, IntegerArray_msg, &dest)))
+        TEST((s = S("\x08\x01"), pb_decode(&s, IntegerArray_msg, &dest)))
+        TEST((s = S("\x08\x01\x00"), pb_decode(&s, IntegerArray_msg, &dest)))
+        TEST((s = S("\x08"), !pb_decode(&s, IntegerArray_msg, &dest)))
+    }
+    
+    {
+        pb_istream_t s;
+        PointerContainer dest;
+        
+        COMMENT("Testing pb_decode with pointer fields")
+        
+        memset(&dest, 0, sizeof(dest));
+#ifdef MALLOC_HEADER
+        TEST((s = S("\x0A\x01\x61\x12\x01\x62\x1A\x00\x2A\x01\x65\x32\x01\x66"
+                  "\x3A\x00\x42\x01\x63\x4A\x01\x64"),
+              pb_decode(&s, PointerContainer_msg, &dest)))
+        TEST(0 == strcmp(dest.text, "a"))
+        TEST(dest.blob.size == 1 && dest.blob.bytes[0] == 'b')
+        TEST(dest.submsg != NULL && dest.submsg->data == 10)
+        TEST(dest.rtext_count == 1 && (0 == strcmp(dest.rtext[0], "e")))
+        TEST(dest.rblob_count == 1 && dest.rblob[0].size == 1 &&
+             dest.rblob[0].bytes[0] == 'f')
+        TEST(dest.rsubmsg_count == 1)
+        TEST(0 == strcmp(dest.otext, "c"))
+        TEST(dest.oblob.size == 1 && dest.oblob.bytes[0] == 'd')
+        TEST(pb_clean(PointerContainer_msg, &dest));
+#else
+        TEST((s = S("\x0A\x01\x61\x12\x01\x62\x2A\x01\x65\x32\x01\x66\x01\x66"
+                  "\x3A\x00\x42\x01\x63\x4A\x01\x64"),
+              !pb_decode(&s, PointerContainer_msg, &dest)))
+#endif
     }
     
     if (status != 0)

tests/encode_unittests.c

@@ -1,6 +1,7 @@
 #include <stdio.h>
 #include <string.h>
 #include "pb_encode.h"
+#include "pb_encode_buffer.h"
 #include "unittests.h"
 #include "unittestproto.pb.h"
 
@@ -23,6 +24,14 @@ bool fieldcallback(pb_ostream_t *stream, const pb_field_t *field, const void *ar
     return pb_encode_varint(stream, value);
 }
 
+bool fieldcallback_buf(pb_strstream_t *stream, const pb_field_t *field, const void *arg)
+{
+    int value = 0x55;
+    if (!pb_encbuf_varint(stream, value))
+        return false;
+    return pb_encbuf_tag_for_field(stream, field);
+}
+
 bool crazyfieldcallback(pb_ostream_t *stream, const pb_field_t *field, const void *arg)
 {
     /* This callback writes different amount of data the second time. */
@@ -33,6 +42,16 @@ bool crazyfieldcallback(pb_ostream_t *stream, const pb_field_t *field, const voi
     return pb_encode_varint(stream, *state);
 }
 
+bool crazyfieldcallback_buf(pb_strstream_t *stream, const pb_field_t *field, const void *arg)
+{
+    /* This callback writes different amount of data the second time. */
+    uint32_t *state = (uint32_t*)arg;
+    *state <<= 8;
+    if (!pb_encbuf_varint(stream, *state))
+        return false;
+    return pb_encbuf_tag_for_field(stream, field);
+}
+
 /* Check that expression x writes data y.
  * Y is a string, which may contain null bytes. Null terminator is ignored.
  */
@@ -43,6 +62,20 @@ s = pb_ostream_from_buffer(buffer, sizeof(buffer)), \
 memcmp(buffer, y, sizeof(y) - 1) == 0 && \
 buffer[sizeof(y) - 1] == 0xAA
 
+/* Check that expression x writes data y into s2.
+ * Y is a string, which may contain null bytes. Null terminator is ignored.
+ */
+#define WRITES_BUF(x, y) \
+memset(buffer, 0xAA, sizeof(buffer)), \
+s2 = pb_strstream_from_buffer(buffer, sizeof(buffer)), \
+(x) && \
+memcmp(s2.last, y, sizeof(y) - 1) == 0 && \
+s2.last - buffer == sizeof(buffer) - (sizeof(y) - 1)
+
+/* Check that expression x and y write data z into s and s2, respectively.
+ */
+#define WRITES_BOTH(x, y, z) (WRITES(x, z)) && (WRITES_BUF(y, z))
+
 int main()
 {
     int status = 0;
@@ -72,51 +105,74 @@ int main()
     {
         uint8_t buffer[30];
         pb_ostream_t s;
+        pb_strstream_t s2;
         
         COMMENT("Test pb_encode_varint")
-        TEST(WRITES(pb_encode_varint(&s, 0), "\0"));
-        TEST(WRITES(pb_encode_varint(&s, 1), "\1"));
-        TEST(WRITES(pb_encode_varint(&s, 0x7F), "\x7F"));
-        TEST(WRITES(pb_encode_varint(&s, 0x80), "\x80\x01"));
-        TEST(WRITES(pb_encode_varint(&s, UINT32_MAX), "\xFF\xFF\xFF\xFF\x0F"));
-        TEST(WRITES(pb_encode_varint(&s, UINT64_MAX), "\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\x01"));
+        TEST(WRITES_BOTH(pb_encode_varint(&s, 0),
+                         pb_encbuf_varint(&s2, 0), "\0"));
+        TEST(WRITES_BOTH(pb_encode_varint(&s, 1),
+                         pb_encbuf_varint(&s2, 1), "\1"));
+        TEST(WRITES_BOTH(pb_encode_varint(&s, 0x7F),
+                         pb_encbuf_varint(&s2, 0x7F), "\x7F"));
+        TEST(WRITES_BOTH(pb_encode_varint(&s, 0x80),
+                         pb_encbuf_varint(&s2, 0x80), "\x80\x01"));
+        TEST(WRITES_BOTH(pb_encode_varint(&s, UINT32_MAX),
+                         pb_encbuf_varint(&s2, UINT32_MAX),
+                         "\xFF\xFF\xFF\xFF\x0F"));
+        TEST(WRITES_BOTH(pb_encode_varint(&s, UINT64_MAX),
+                         pb_encbuf_varint(&s2, UINT64_MAX),
+                         "\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\x01"));
     }
     
     {
         uint8_t buffer[30];
         pb_ostream_t s;
+        pb_strstream_t s2;
         
         COMMENT("Test pb_encode_tag")
-        TEST(WRITES(pb_encode_tag(&s, PB_WT_STRING, 5), "\x2A"));
-        TEST(WRITES(pb_encode_tag(&s, PB_WT_VARINT, 99), "\x98\x06"));
+        TEST(WRITES_BOTH(pb_encode_tag(&s, PB_WT_STRING, 5),
+                         pb_encbuf_tag(&s2, PB_WT_STRING, 5), "\x2A"));
+        TEST(WRITES_BOTH(pb_encode_tag(&s, PB_WT_VARINT, 99),
+                         pb_encbuf_tag(&s2, PB_WT_VARINT, 99), "\x98\x06"));
     }
     
     {
         uint8_t buffer[30];
         pb_ostream_t s;
+        pb_strstream_t s2;
         pb_field_t field = {10, PB_LTYPE_SVARINT};
         
         COMMENT("Test pb_encode_tag_for_field")
-        TEST(WRITES(pb_encode_tag_for_field(&s, &field), "\x50"));
+        TEST(WRITES_BOTH(pb_encode_tag_for_field(&s, &field),
+                         pb_encbuf_tag_for_field(&s2, &field), "\x50"));
         
         field.type = PB_LTYPE_FIXED64;
-        TEST(WRITES(pb_encode_tag_for_field(&s, &field), "\x51"));
+        TEST(WRITES_BOTH(pb_encode_tag_for_field(&s, &field),
+                         pb_encbuf_tag_for_field(&s2, &field), "\x51"));
         
         field.type = PB_LTYPE_STRING;
-        TEST(WRITES(pb_encode_tag_for_field(&s, &field), "\x52"));
+        TEST(WRITES_BOTH(pb_encode_tag_for_field(&s, &field),
+                         pb_encbuf_tag_for_field(&s2, &field), "\x52"));
         
         field.type = PB_LTYPE_FIXED32;
-        TEST(WRITES(pb_encode_tag_for_field(&s, &field), "\x55"));
+        TEST(WRITES_BOTH(pb_encode_tag_for_field(&s, &field),
+                         pb_encbuf_tag_for_field(&s2, &field), "\x55"));
     }
     
     {
         uint8_t buffer[30];
         pb_ostream_t s;
+        pb_strstream_t s2;
         
         COMMENT("Test pb_encode_string")
-        TEST(WRITES(pb_encode_string(&s, (const uint8_t*)"abcd", 4), "\x04""abcd"));
-        TEST(WRITES(pb_encode_string(&s, (const uint8_t*)"abcd\x00", 5), "\x05""abcd\x00"));
-        TEST(WRITES(pb_encode_string(&s, (const uint8_t*)"", 0), "\x00"));
+        TEST(WRITES_BOTH(pb_encode_string(&s, (const uint8_t*)"abcd", 4),
+                         pb_encbuf_string(&s2, (const uint8_t*)"abcd", 4),
+                         "\x04""abcd"));
+        TEST(WRITES_BOTH(pb_encode_string(&s, (const uint8_t*)"abcd\x00", 5),
+                         pb_encbuf_string(&s2, (const uint8_t*)"abcd\x00", 5),
+                         "\x05""abcd\x00"));
+        TEST(WRITES_BOTH(pb_encode_string(&s, (const uint8_t*)"", 0),
+                         pb_encbuf_string(&s2, (const uint8_t*)"", 0), "\x00"));
     }
     
     {
@@ -192,81 +248,154 @@ int main()
     {
         uint8_t buffer[10];
         pb_ostream_t s;
-        IntegerArray msg = {5, {1, 2, 3, 4, 5}};
+        pb_strstream_t s2;
+        IntegerArray msg = {{0}, 5, {1, 2, 3, 4, 5}};
         
         COMMENT("Test pb_encode with int32 array")
         
-        TEST(WRITES(pb_encode(&s, IntegerArray_fields, &msg), "\x0A\x05\x01\x02\x03\x04\x05"))
+        TEST(WRITES_BOTH(pb_encode(&s, IntegerArray_msg, &msg),
+                         pb_encode_buffer(&s2, IntegerArray_msg, &msg),
+                         "\x0A\x05\x01\x02\x03\x04\x05"))
         
         msg.data_count = 0;
-        TEST(WRITES(pb_encode(&s, IntegerArray_fields, &msg), ""))
+        TEST(WRITES_BOTH(pb_encode(&s, IntegerArray_msg, &msg),
+                         pb_encode_buffer(&s2, IntegerArray_msg, &msg), ""))
         
         msg.data_count = 10;
-        TEST(!pb_encode(&s, IntegerArray_fields, &msg))
+        TEST(!pb_encode(&s, IntegerArray_msg, &msg))
+        TEST(!pb_encode_buffer(&s2, IntegerArray_msg, &msg))
     }
     
     {
         uint8_t buffer[10];
         pb_ostream_t s;
-        FloatArray msg = {1, {99.0f}};
+        pb_strstream_t s2;
+        FloatArray msg = {{0}, 1, {99.0f}};
         
         COMMENT("Test pb_encode with float array")
         
-        TEST(WRITES(pb_encode(&s, FloatArray_fields, &msg),
-                    "\x0A\x04\x00\x00\xc6\x42"))
+        TEST(WRITES_BOTH(pb_encode(&s, FloatArray_msg, &msg),
+                         pb_encode_buffer(&s2, FloatArray_msg, &msg),
+                         "\x0A\x04\x00\x00\xc6\x42"))
         
         msg.data_count = 0;
-        TEST(WRITES(pb_encode(&s, FloatArray_fields, &msg), ""))
+        TEST(WRITES_BOTH(pb_encode(&s, FloatArray_msg, &msg),
+                         pb_encode_buffer(&s2, FloatArray_msg, &msg), ""))
         
         msg.data_count = 3;
-        TEST(!pb_encode(&s, FloatArray_fields, &msg))
+        TEST(!pb_encode(&s, FloatArray_msg, &msg))
+        TEST(!pb_encode_buffer(&s2, FloatArray_msg, &msg))
     }
     
     {
         uint8_t buffer[10];
         pb_ostream_t s;
+        pb_strstream_t s2;
         CallbackArray msg;
         
-        msg.data.funcs.encode = &fieldcallback;
-        
         COMMENT("Test pb_encode with callback field.")
-        TEST(WRITES(pb_encode(&s, CallbackArray_fields, &msg), "\x08\x55"))
+        msg.data.funcs.encode = &fieldcallback;
+        TEST(WRITES(pb_encode(&s, CallbackArray_msg, &msg), "\x08\x55"))
+        msg.data.funcs.encode_buffer = &fieldcallback_buf;
+        TEST(WRITES_BUF(pb_encode_buffer(&s2, CallbackArray_msg, &msg), "\x08\x55"))
     }
     
     {
         uint8_t buffer[10];
         pb_ostream_t s;
-        IntegerContainer msg = {{5, {1,2,3,4,5}}};
+        pb_strstream_t s2;
+        IntegerContainer msg = {{0}, {{0}, 5, {1,2,3,4,5}}};
         
         COMMENT("Test pb_encode with packed array in a submessage.")
-        TEST(WRITES(pb_encode(&s, IntegerContainer_fields, &msg),
-                    "\x0A\x07\x0A\x05\x01\x02\x03\x04\x05"))
+        TEST(WRITES_BOTH(pb_encode(&s, IntegerContainer_msg, &msg),
+                         pb_encode_buffer(&s2, IntegerContainer_msg, &msg),
+                         "\x0A\x07\x0A\x05\x01\x02\x03\x04\x05"))
     }
     
     {
         uint8_t buffer[10];
         pb_ostream_t s;
+        pb_strstream_t s2;
         CallbackContainer msg;
         CallbackContainerContainer msg2;
         uint32_t state = 1;
         
+        COMMENT("Test pb_encode with callback field in a submessage.")
         msg.submsg.data.funcs.encode = &fieldcallback;
         msg2.submsg.submsg.data.funcs.encode = &fieldcallback;
-        
-        COMMENT("Test pb_encode with callback field in a submessage.")
-        TEST(WRITES(pb_encode(&s, CallbackContainer_fields, &msg), "\x0A\x02\x08\x55"))
-        TEST(WRITES(pb_encode(&s, CallbackContainerContainer_fields, &msg2),
+        TEST(WRITES(pb_encode(&s, CallbackContainer_msg, &msg), "\x0A\x02\x08\x55"))
+        TEST(WRITES(pb_encode(&s, CallbackContainerContainer_msg, &msg2),
                     "\x0A\x04\x0A\x02\x08\x55"))
         
+        msg.submsg.data.funcs.encode_buffer = &fieldcallback_buf;
+        msg2.submsg.submsg.data.funcs.encode_buffer = &fieldcallback_buf;
+        TEST(WRITES_BUF(pb_encode_buffer(&s2, CallbackContainer_msg, &msg), "\x0A\x02\x08\x55"))
+        TEST(WRITES_BUF(pb_encode_buffer(&s2, CallbackContainerContainer_msg, &msg2),
+                        "\x0A\x04\x0A\x02\x08\x55"))
+        
         /* Misbehaving callback: varying output between calls */
         msg.submsg.data.funcs.encode = &crazyfieldcallback;
         msg.submsg.data.arg = &state;
         msg2.submsg.submsg.data.funcs.encode = &crazyfieldcallback;
         msg2.submsg.submsg.data.arg = &state;
         
-        TEST(!pb_encode(&s, CallbackContainer_fields, &msg))
+        TEST(!pb_encode(&s, CallbackContainer_msg, &msg))
         state = 1;
-        TEST(!pb_encode(&s, CallbackContainerContainer_fields, &msg2))
+        TEST(!pb_encode(&s, CallbackContainerContainer_msg, &msg2))
+
+        msg.submsg.data.funcs.encode_buffer = &crazyfieldcallback_buf;
+        msg.submsg.data.arg = &state;
+        msg2.submsg.submsg.data.funcs.encode_buffer = &crazyfieldcallback_buf;
+        msg2.submsg.submsg.data.arg = &state;
+        
+        TEST(!pb_encode_buffer(&s2, CallbackContainer_msg, &msg))
+        state = 1;
+        TEST(!pb_encode_buffer(&s2, CallbackContainerContainer_msg, &msg2))
+    }
+    
+    {
+        uint8_t buffer[30];
+        pb_ostream_t s;
+        pb_strstream_t s2;
+        PointerContainer msg;
+        DefaultContainer msg2;
+        IntegerArray msg3;
+        
+        COMMENT("Test pb_encode with pointer fields.")
+        
+        memset(&msg, 0, sizeof(msg));
+        memset(&msg2, 0, sizeof(msg2));
+        msg.text = "a";
+        msg.blob.size = 1;
+        msg.blob.bytes = (uint8_t*)"b";
+        msg.submsg = &msg2;
+        TEST(WRITES_BOTH(pb_encode(&s, PointerContainer_msg, &msg),
+                         pb_encode_buffer(&s2, PointerContainer_msg, &msg),
+                         "\x0A\x01\x61\x12\x01\x62\x1A\x00"))
+        
+        memset(&msg3, 0, sizeof(msg3));
+        msg.rtext_count = 1;
+        msg.rtext[0] = "e";
+        msg.rblob_count = 1;
+        msg.rblob[0].size = 1;
+        msg.rblob[0].bytes = (uint8_t*)"f";
+        msg.rsubmsg_count = 1;
+        msg.rsubmsg[0] = &msg3;
+        TEST(WRITES_BOTH(pb_encode(&s, PointerContainer_msg, &msg),
+                         pb_encode_buffer(&s2, PointerContainer_msg, &msg),
+                         "\x0A\x01\x61\x12\x01\x62\x1A\x00"
+                         "\x2A\x01\x65\x32\x01\x66\x3A\x00"));
+        
+        PointerContainer_set(msg, otext);
+        msg.otext = "c";
+        PointerContainer_set(msg, oblob);
+        msg.oblob.size = 1;
+        msg.oblob.bytes = (uint8_t*)"d";
+        TEST(WRITES_BOTH(pb_encode(&s, PointerContainer_msg, &msg),
+                         pb_encode_buffer(&s2, PointerContainer_msg, &msg),
+                         "\x0A\x01\x61\x12\x01\x62\x1A\x00"
+                         "\x2A\x01\x65\x32\x01\x66\x3A\x00"
+                         "\x42\x01\x63\x4A\x01\x64"));
     }
     
     if (status != 0)

tests/test_decode1.c

@@ -17,7 +17,7 @@ bool print_person(pb_istream_t *stream)
     int i;
     Person person;
     
-    if (!pb_decode(stream, Person_fields, &person))
+    if (!pb_decode(stream, Person_msg, &person))
         return false;
     
     /* Now the decoding is done, rest is just to print stuff out. */
@@ -25,7 +25,7 @@ bool print_person(pb_istream_t *stream)
     printf("name: \"%s\"\n", person.name);
     printf("id: %ld\n", (long)person.id);
     
-    if (person.has_email)
+    if (Person_has(person, email))
         printf("email: \"%s\"\n", person.email);
     
     for (i = 0; i < person.phone_count; i++)
@@ -34,7 +34,7 @@ bool print_person(pb_istream_t *stream)
         printf("phone {\n");
         printf("  number: \"%s\"\n", phone->number);
         
-        if (phone->has_type)
+        if (Person_PhoneNumber_has(*phone, type))
         {
             switch (phone->type)
             {

tests/test_decode2.c

@@ -13,7 +13,7 @@ bool print_person(pb_istream_t *stream)
     int i;
     Person person;
     
-    if (!pb_decode(stream, Person_fields, &person))
+    if (!pb_decode(stream, Person_msg, &person))
         return false;
     
     /* Now the decoding is done, rest is just to print stuff out. */
@@ -21,7 +21,7 @@ bool print_person(pb_istream_t *stream)
     printf("name: \"%s\"\n", person.name);
     printf("id: %ld\n", (long)person.id);
     
-    if (person.has_email)
+    if (Person_has(person, email))
         printf("email: \"%s\"\n", person.email);
     
     for (i = 0; i < person.phone_count; i++)
@@ -30,7 +30,7 @@ bool print_person(pb_istream_t *stream)
         printf("phone {\n");
         printf("  number: \"%s\"\n", phone->number);
         
-        if (phone->has_type)
+        if (Person_PhoneNumber_has(*phone, type))
         {
             switch (phone->type)
             {

tests/test_decode_callbacks.c

@@ -84,7 +84,7 @@ int main()
     testmessage.fixed64value.funcs.decode = &print_fixed64;
     testmessage.fixed64value.arg = "fixed64value: %lld\n";
     
-    if (!pb_decode(&stream, TestMessage_fields, &testmessage))
+    if (!pb_decode(&stream, TestMessage_msg, &testmessage))
         return 1;
     
     return 0;

tests/test_encode1.c

@@ -10,17 +10,21 @@
 int main()
 {
     /* Initialize the structure with constants */
-    Person person = {"Test Person 99", 99, true, "test@person.com",
-        3, {{"555-12345678", true, Person_PhoneType_MOBILE},
-            {"99-2342", false, 0},
-            {"1234-5678", true, Person_PhoneType_WORK},
+    Person person = {{1 << Person_email_index},
+        "Test Person 99", 99, "test@person.com",
+        3, {{{1 << Person_PhoneNumber_type_index},
+             "555-12345678", Person_PhoneType_MOBILE},
+            {{1 << Person_PhoneNumber_type_index},
+             "99-2342", 0},
+            {{1 << Person_PhoneNumber_type_index},
+             "1234-5678", Person_PhoneType_WORK},
         }};
 
     uint8_t buffer[512];
     pb_ostream_t stream = pb_ostream_from_buffer(buffer, sizeof(buffer));
     
     /* Now encode it and check if we succeeded. */
-    if (pb_encode(&stream, Person_fields, &person))
+    if (pb_encode(&stream, Person_msg, &person))
     {
         fwrite(buffer, 1, stream.bytes_written, stdout);
         return 0; /* Success */

tests/test_encode2.c

@@ -15,17 +15,21 @@ bool streamcallback(pb_ostream_t *stream, const uint8_t *buf, size_t count)
 int main()
 {
     /* Initialize the structure with constants */
-    Person person = {"Test Person 99", 99, true, "test@person.com",
-        3, {{"555-12345678", true, Person_PhoneType_MOBILE},
-            {"99-2342", false, 0},
-            {"1234-5678", true, Person_PhoneType_WORK},
+    Person person = {{1 << Person_email_index},
+        "Test Person 99", 99, "test@person.com",
+        3, {{{1 << Person_PhoneNumber_type_index},
+             "555-12345678", Person_PhoneType_MOBILE},
+            {{1 << Person_PhoneNumber_type_index},
+             "99-2342", 0},
+            {{1 << Person_PhoneNumber_type_index},
+             "1234-5678", Person_PhoneType_WORK},
         }};
     
     /* Prepare the stream, output goes directly to stdout */
     pb_ostream_t stream = {&streamcallback, stdout, SIZE_MAX, 0};
     
     /* Now encode it and check if we succeeded. */
-    if (pb_encode(&stream, Person_fields, &person))
+    if (pb_encode(&stream, Person_msg, &person))
         return 0; /* Success */
     else
         return 1; /* Failure */

tests/test_encode_callbacks.c

@@ -52,13 +52,13 @@ int main()
     testmessage.fixed32value.funcs.encode = &encode_fixed32;
     testmessage.fixed64value.funcs.encode = &encode_fixed64;
     
-    testmessage.has_submsg = true;
+    TestMessage_set(testmessage, submsg);
     testmessage.submsg.stringvalue.funcs.encode = &encode_string;
     testmessage.submsg.int32value.funcs.encode = &encode_int32;
     testmessage.submsg.fixed32value.funcs.encode = &encode_fixed32;
     testmessage.submsg.fixed64value.funcs.encode = &encode_fixed64;
     
-    if (!pb_encode(&stream, TestMessage_fields, &testmessage))
+    if (!pb_encode(&stream, TestMessage_msg, &testmessage))
         return 1;
     
     if (fwrite(buffer, stream.bytes_written, 1, stdout) != 1)

tests/unittestproto.proto

@@ -26,3 +26,28 @@ message CallbackContainer {
 message CallbackContainerContainer {
     required CallbackContainer submsg = 1;
 }
+
+message DefaultContainer {
+    optional int32 data = 1 [default = 10];
+}
+
+message PointerContainer {
+    required string text = 1 [(nanopb).pointer = true];
+    required bytes blob = 2 [(nanopb).pointer = true];
+    required DefaultContainer submsg = 3 [(nanopb).pointer = true];
+    // This should be rejected:
+    // required int32 data = 4 [(nanopb).pointer = true];
+    repeated string rtext = 5 [(nanopb).pointer = true, (nanopb).max_count = 10];
+    repeated bytes rblob = 6 [(nanopb).pointer = true, (nanopb).max_count = 10];
+    repeated IntegerArray rsubmsg = 7 [(nanopb).pointer = true, (nanopb).max_count = 10];
+    optional string otext = 8 [(nanopb).pointer = true];
+    optional bytes oblob = 9 [(nanopb).pointer = true];
+}
+
+message RecursiveRef_A {
+    optional RecursiveRef_B submsg = 1 [(nanopb).pointer = true];
+}
+
+message RecursiveRef_B {
+    required RecursiveRef_A submsg = 1;
+}