First version of decoding

git-svn-id: https://svn.kapsi.fi/jpa/nanopb@942 e3a754e5-d11d-0410-8d38-ebb782a927b9

6 files changed, 559 insertions, 0 deletions

diff --git a/pb.h b/pb.h
new file mode 100644
index 00000000..2cae75c5
--- /dev/null
+++ b/pb.h
@@ -0,0 +1,107 @@
+#ifndef _PB_H_
+#define _PB_H_
+
+#include <stdint.h>
+#include <stddef.h> // size_t
+
+/* Lightweight input stream.
+ * If buf is NULL, read but don't store bytes. */
+typedef struct _pb_istream_t pb_istream_t;
+struct _pb_istream_t
+{
+    bool (*callback)(pb_istream_t *stream, char *buf, size_t count);
+    void *state; // Free field for use by callback implementation
+    size_t bytes_left;
+};
+
+static inline bool pb_read(pb_istream_t *stream, char *buf, size_t count)
+{
+    bool status = stream->callback(stream, buf, count);
+    stream->bytes_left -= count;
+    return status;
+}
+
+/* Lightweight output stream. */
+typedef struct _pb_ostream_t pb_ostream_t;
+struct _pb_ostream_t
+{
+    bool (*callback)(pb_ostream_t *stream, const char *buf, size_t count);
+    void *state; // Free field for use by callback implementation
+    size_t bytes_written;
+};
+
+static inline bool pb_write(pb_ostream_t *stream, const char *buf, size_t count)
+{
+    bool status = stream->callback(stream, buf, count);
+    stream->bytes_written += count;
+    return status;
+}
+
+/* List of possible decode/encode action types */
+
+typedef enum {
+    // Special case. Sets boolean field to true, continues parsing the data.
+    PB_ACT_HAS,
+    
+    // Standard integer types
+    PB_ACT_UINT32,
+    PB_ACT_SINT32,
+    PB_ACT_INT32,
+    PB_ACT_FIXED32,
+    PB_ACT_SFIXED32,
+    PB_ACT_UINT64,
+    PB_ACT_SINT64,
+    PB_ACT_INT64,
+    PB_ACT_FIXED64,
+    PB_ACT_SFIXED64,
+    PB_ACT_BOOL,
+    
+    // Standard float types
+    PB_ACT_FLOAT,
+    PB_ACT_DOUBLE,
+    
+    // Constant-sized array
+    PB_ACT_BYTES,
+    
+    // Constant-sized array, with null termination
+    PB_ACT_STRING,
+    
+    // Callback function pointer in field value
+    PB_ACT_SUBMESSAGE,
+    
+    PB_LAST_ACT
+} pb_action_t;
+
+// This structure is used in constants to specify struct fields.
+typedef struct {
+    int field_number;
+    uint16_t offset;
+    pb_action_t action;
+    uint8_t fieldsize;
+} pb_field_t;
+
+#define PB_LAST_FIELD {0,0,0,0}
+
+/* --- Types to use inside generated structures. --- */
+
+// Byte array and size word.
+// Note: because of variable length array, this type cannot be directly used.
+// Autogenerated code declares the same type of fields but with explicit length.
+typedef struct {
+    size_t size;
+    char bytes[];
+} pb_bytearray_t;
+
+// This structure is used for giving the callback function.
+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, void *arg);
+    } funcs;
+    
+    // Free arg for use by callback
+    void *arg;
+};
+
+#endif
\ No newline at end of file
diff --git a/pb_decode.c b/pb_decode.c
new file mode 100644
index 00000000..884cf18e
--- /dev/null
+++ b/pb_decode.c
@@ -0,0 +1,274 @@
+/* pb_decode.c -- decode a protobuf using callback functions
+ *
+ * 2011 Petteri Aimonen <jpa@kapsi.fi>
+ */
+
+#include "pb_decode.h"
+
+const pb_decoder_t PB_DECODERS[PB_LAST_ACT] = {
+    NULL,
+    &pb_dec_uint32,
+    &pb_dec_sint32,
+    &pb_dec_uint32, // Cast to int32
+    &pb_dec_fixed32,
+    &pb_dec_fixed32, // Cast to int32
+    &pb_dec_uint64,
+    &pb_dec_sint64,
+    &pb_dec_uint64, // Cast to int64
+    &pb_dec_fixed64,
+    &pb_dec_fixed64, // Cast to int64
+    &pb_dec_bool,
+    &pb_dec_float,
+    &pb_dec_double,
+    &pb_dec_bytes,
+    &pb_dec_string,
+    &pb_dec_submessage
+};
+
+enum wire_type {
+    WT_VARINT = 0,
+    WT_64BIT  = 1,
+    WT_STRING = 2,
+    WT_32BIT  = 5
+};
+
+// Note: pb_decode_varint32 is a bit un-orthodox:
+// it will refuse to decode values that exceed uint32 range.
+// The Google implementation would simply cast to 32 bits.
+bool pb_decode_varint32(pb_istream_t *stream, uint32_t *dest)
+{
+    char byte;
+    int bitpos = 0;
+    *dest = 0;
+    
+    while (bitpos < 32 && pb_read(stream, &byte, 1))
+    {
+        *dest |= (byte & 0x7F) << bitpos;
+        bitpos += 7;
+        
+        if (!(byte & 0x80))
+            return true;
+    }
+    
+    return false;
+}
+
+bool pb_decode_varint64(pb_istream_t *stream, uint64_t *dest)
+{
+    char byte;
+    int bitpos = 0;
+    *dest = 0;
+    
+    while (bitpos < 64 && pb_read(stream, &byte, 1))
+    {
+        *dest |= (byte & 0x7F) << bitpos;
+        bitpos += 7;
+        
+        if (!(byte & 0x80))
+            return true;
+    }
+    
+    return false;
+}
+
+bool pb_skip_varint(pb_istream_t *stream)
+{
+    char byte;
+    do
+    {
+        if (!pb_read(stream, &byte, 1))
+            return false;
+    } while (byte & 0x80);
+    return true;
+}
+
+bool pb_skip_string(pb_istream_t *stream)
+{
+    uint32_t length;
+    if (!pb_decode_varint32(stream, &length))
+        return false;
+    
+    return pb_read(stream, NULL, length);
+}
+
+bool pb_decode(pb_istream_t *stream, const pb_field_t fields[], void *dest)
+{
+    while (stream->bytes_left)
+    {
+        uint32_t temp;
+        if (!pb_decode_varint32(stream, &temp))
+            return false;
+        
+        int field_number = temp >> 3;
+        int wire_type = temp & 7;
+        
+        const pb_field_t *field = fields;
+        while (field->field_number != 0)
+        {
+            if (field->field_number != field_number)
+            {
+                field++;
+                continue;
+            }
+            
+            void *destfield = dest + field->offset; 
+            
+            if (field->action == PB_ACT_HAS)
+            {
+                *(bool*)destfield = true;
+                field++;
+                continue;
+            }
+            
+            pb_decoder_t func = PB_DECODERS[field->action];
+            if (!func(stream, field, destfield))
+                return false;
+            
+            break;
+        }
+        
+        if (field->field_number == 0) // No match found, skip data
+        {
+            bool status = false;
+            switch (wire_type)
+            {
+                case WT_VARINT:
+                    status = pb_skip_varint(stream);
+                    break;
+                
+                case WT_64BIT:
+                    status = pb_read(stream, NULL, 8);
+                    break;
+                
+                case WT_STRING:
+                    status = pb_skip_string(stream);
+                    break;
+                
+                case WT_32BIT:
+                    status = pb_read(stream, NULL, 4);
+                    break;
+            }
+            
+            if (!status)
+                return false;
+        }
+    }
+    
+    return true;
+}
+
+bool pb_dec_uint32(pb_istream_t *stream, const pb_field_t *field, void *dest)
+{
+    return pb_decode_varint32(stream, (uint32_t*)dest);
+}
+
+bool pb_dec_sint32(pb_istream_t *stream, const pb_field_t *field, void *dest)
+{
+    uint32_t *x = (uint32_t*)dest;
+    bool status = pb_decode_varint32(stream, x);
+    *x = (*x >> 1) ^ -(int32_t)(*x & 1);
+    return status;
+}
+
+bool pb_dec_fixed32(pb_istream_t *stream, const pb_field_t *field, void *dest)
+{
+    char bytes[4] = {0};
+    bool status = pb_read(stream, bytes, 4);
+    *(uint32_t*)dest = 
+        bytes[0] | ((uint32_t)bytes[1] << 8) |
+        ((uint32_t)bytes[2] << 16) | ((uint32_t)bytes[3] << 24);
+    return status;
+}
+
+bool pb_dec_uint64(pb_istream_t *stream, const pb_field_t *field, void *dest)
+{
+    return pb_decode_varint64(stream, (uint64_t*)dest);
+}
+
+bool pb_dec_sint64(pb_istream_t *stream, const pb_field_t *field, void *dest)
+{
+    uint64_t *x = (uint64_t*)dest;
+    bool status = pb_decode_varint64(stream, x);
+    *x = (*x >> 1) ^ -(int64_t)(*x & 1);
+    return status;
+}
+
+bool pb_dec_fixed64(pb_istream_t *stream, const pb_field_t *field, void *dest)
+{
+    char bytes[8] = {0};
+    bool status = pb_read(stream, bytes, 8);
+    *(uint64_t*)dest =
+        (uint64_t)bytes[0] | ((uint64_t)bytes[1] << 8) |
+        ((uint64_t)bytes[2] << 16) | ((uint64_t)bytes[3] << 24) |
+        ((uint64_t)bytes[4] << 32) | ((uint64_t)bytes[5] << 40) |
+        ((uint64_t)bytes[6] << 48) | ((uint64_t)bytes[7] << 56);
+    return status;
+}
+
+bool pb_dec_bool(pb_istream_t *stream, const pb_field_t *field, void *dest)
+{
+    uint32_t temp = 0;
+    bool status = pb_decode_varint32(stream, &temp);
+    *(bool*)dest = !!temp;
+    return status;
+}
+
+bool pb_dec_float(pb_istream_t *stream, const pb_field_t *field, void *dest)
+{
+    return pb_read(stream, (char*)dest, sizeof(float));
+}
+
+bool pb_dec_double(pb_istream_t *stream, const pb_field_t *field, void *dest)
+{
+    return pb_read(stream, (char*)dest, sizeof(double));
+}
+
+bool pb_dec_bytes(pb_istream_t *stream, const pb_field_t *field, void *dest)
+{
+    pb_bytearray_t *x = (pb_bytearray_t*)dest;
+    uint32_t temp;
+    if (!pb_decode_varint32(stream, &temp))
+        return false;
+    x->size = temp;
+    
+    if (x->size > field->fieldsize)
+        return false;
+    
+    return pb_read(stream, x->bytes, x->size);
+}
+
+bool pb_dec_string(pb_istream_t *stream, const pb_field_t *field, void *dest)
+{
+    uint32_t size;
+    if (!pb_decode_varint32(stream, &size))
+        return false;
+    
+    if (size > field->fieldsize - 1)
+        return false;
+    
+    bool status = pb_read(stream, (char*)dest, size);
+    *((char*)dest + size) = 0;
+    return status;
+}
+
+bool pb_dec_submessage(pb_istream_t *stream, const pb_field_t *field, void *dest)
+{
+    pb_callback_t *x = (pb_callback_t*)dest;
+    
+    if (x->funcs.decode == NULL)
+        return pb_skip_string(stream);
+    
+    uint32_t size;
+    if (!pb_decode_varint32(stream, &size))
+        return false;
+    
+    if (stream->bytes_left < size)
+        return false;
+    
+    // Make a limited-length istream for decoding submessage
+    pb_istream_t shortstream = *stream;
+    shortstream.bytes_left = size;
+    bool status = x->funcs.decode(&shortstream, field, x->arg);
+    stream->bytes_left -= size - shortstream.bytes_left;
+    return status;
+}
diff --git a/pb_decode.h b/pb_decode.h
new file mode 100644
index 00000000..b71e20b1
--- /dev/null
+++ b/pb_decode.h
@@ -0,0 +1,55 @@
+#ifndef _PB_DECODE_H_
+#define _PB_DECODE_H_
+
+#include <stdbool.h>
+#include "pb.h"
+
+// Decode from stream to destination struct.
+// The actual struct pointed to by dest must match the description in fields.
+bool pb_decode(pb_istream_t *stream, const pb_field_t fields[], void *dest);
+
+/* --- Helper functions ---
+ * You may want to use these from your callbacks.
+ */
+
+bool pb_decode_varint32(pb_istream_t *stream, uint32_t *dest);
+bool pb_decode_varint64(pb_istream_t *stream, uint64_t *dest);
+
+bool pb_skip_varint(pb_istream_t *stream);
+bool pb_skip_string(pb_istream_t *stream);
+
+/* --- Field decoders ---
+ * Each decoder takes stream and field description, and a pointer to the field
+ * in the destination struct (dest = struct_addr + field->offset).
+ */
+
+// Integer types.
+bool pb_dec_uint32(pb_istream_t *stream, const pb_field_t *field, void *dest);
+bool pb_dec_sint32(pb_istream_t *stream, const pb_field_t *field, void *dest);
+bool pb_dec_fixed32(pb_istream_t *stream, const pb_field_t *field, void *dest);
+bool pb_dec_uint64(pb_istream_t *stream, const pb_field_t *field, void *dest);
+bool pb_dec_sint64(pb_istream_t *stream, const pb_field_t *field, void *dest);
+bool pb_dec_fixed64(pb_istream_t *stream, const pb_field_t *field, void *dest);
+bool pb_dec_bool(pb_istream_t *stream, const pb_field_t *field, void *dest);
+
+// Floating point types. Info is ignored.
+bool pb_dec_float(pb_istream_t *stream, const pb_field_t *field, void *dest);
+bool pb_dec_double(pb_istream_t *stream, const pb_field_t *field, void *dest);
+
+// Byte array. Dest is pointer to 
+bool pb_dec_bytes(pb_istream_t *stream, const pb_field_t *field, void *dest);
+
+// Null-terminated string.
+bool pb_dec_string(pb_istream_t *stream, const pb_field_t *field, void *dest);
+
+// Use callback. Dest is pointer to pb_callback_t struct.
+bool pb_dec_submessage(pb_istream_t *stream, const pb_field_t *field, void *dest);
+
+typedef bool (*pb_decoder_t)(pb_istream_t *stream, const pb_field_t *field, void *dest);
+
+/* --- Function pointers to field decoders ---
+ * Order in the array must match pb_action_t numbering.
+ */
+const pb_decoder_t PB_DECODERS[PB_LAST_ACT];
+
+#endif
diff --git a/tests/person.proto b/tests/person.proto
new file mode 100644
index 00000000..ec5b4cca
--- /dev/null
+++ b/tests/person.proto
@@ -0,0 +1,18 @@
+message Person {
+  required string name = 1;
+  required int32 id = 2;
+  optional string email = 3;
+
+  enum PhoneType {
+    MOBILE = 0;
+    HOME = 1;
+    WORK = 2;
+  }
+
+  message PhoneNumber {
+    required string number = 1;
+    optional PhoneType type = 2 [default = HOME];
+  }
+
+  repeated PhoneNumber phone = 4;
+}
diff --git a/tests/test_decode1.c b/tests/test_decode1.c
new file mode 100644
index 00000000..dc0831a2
--- /dev/null
+++ b/tests/test_decode1.c
@@ -0,0 +1,102 @@
+#include <stdio.h>
+#include <string.h>
+#include <stddef.h>
+#include "pb_decode.h"
+
+/* Structures for "Person" message */
+
+typedef enum {
+    Person_PhoneType_MOBILE = 0,
+    Person_PhoneType_HOME = 1,
+    Person_PhoneType_WORK = 2,
+    
+    _Person_PhoneType_size = 0xFFFFFFFF // Force 32-bit enum
+} Person_PhoneType;
+
+typedef struct {
+    char number[40];
+    bool has_type;
+    Person_PhoneType type;
+} Person_PhoneNumber;
+
+typedef struct {
+    char name[40];
+    int32_t id;
+    bool has_email;
+    char email[40];
+
+    pb_callback_t phone;
+} Person;
+
+/* Field descriptions */
+#define membersize(st, m) (sizeof ((st*)0)->m)
+
+const pb_field_t Person_PhoneNumber_fields[] = {
+    {1, offsetof(Person_PhoneNumber, number), PB_ACT_STRING, membersize(Person_PhoneNumber, number)},
+    {2, offsetof(Person_PhoneNumber, has_type), PB_ACT_HAS, membersize(Person_PhoneNumber, has_type)},
+    {2, offsetof(Person_PhoneNumber, type), PB_ACT_UINT32, membersize(Person_PhoneNumber, type)},
+    PB_LAST_FIELD
+};
+
+const pb_field_t Person_fields[] = {
+    {1, offsetof(Person, name), PB_ACT_STRING, membersize(Person, name)},
+    {2, offsetof(Person, id), PB_ACT_INT32, membersize(Person, id)},
+    {3, offsetof(Person, email), PB_ACT_STRING, membersize(Person, email)},
+    {4, offsetof(Person, phone), PB_ACT_SUBMESSAGE, membersize(Person, phone)}
+};
+
+/* Default value descriptions */
+#define Person_PhoneNumber_default {"", false, Person_PhoneType_HOME};
+#define Person_default {"", 0, false, "", {{0},0}};
+
+/* And now, the actual test program */
+
+bool print_phonenumber(pb_istream_t *stream, const pb_field_t *field, void *arg)
+{
+    Person_PhoneNumber x = Person_PhoneNumber_default;
+    if (!pb_decode(stream, Person_PhoneNumber_fields, &x))
+        return false;
+    
+    printf("PhoneNumber: number '%s' type '%d'\n", x.number, x.type);
+    return true;
+}
+
+bool print_person(pb_istream_t *stream)
+{
+    Person x = Person_default;
+    x.phone.funcs.decode = &print_phonenumber;
+    
+    if (!pb_decode(stream, Person_fields, &x))
+        return false;
+    
+    printf("Person: name '%s' id '%d' email '%s'\n", x.name, x.id, x.email);
+    return true;
+}
+
+bool my_read(pb_istream_t *stream, char *buf, size_t count)
+{
+    char *source = (char*)stream->state;
+    
+    if (!stream->bytes_left)
+        return false;
+    
+    if (buf != NULL)
+    {
+        memcpy(buf, source, count);
+    }
+    
+    stream->state = source + count;
+    return true;
+}
+
+int main()
+{
+    char buffer[512];
+    size_t size = fread(buffer, 1, 512, stdin);
+    
+    pb_istream_t stream = {&my_read, buffer, size};
+    if (!print_person(&stream))
+        printf("Parsing failed.\n");
+    
+    return 0;
+}
diff --git a/tests/testperson.pb b/tests/testperson.pb
new file mode 100644
index 00000000..d1eb8cfb
--- /dev/null
+++ b/tests/testperson.pb
@@ -0,0 +1,3 @@
+
+Test Person7foobar@foobar.com"
+555-12345678
\ No newline at end of file