polar1225 / protoc-gen-as3 Goto Github PK

I would like to use no dynamic class.
And keep unknown field value.

Best regards

Is there any reason that generated classes are "public final dynamic class" 
rather than just "public final class"? The dynamic keyword essentially turns 
off type-checking in the as3 compiler, which makes working with protobuf 
objects more error-prone.

Is it possible to have as3_package option just like java_package to specify
namespace for the generated as3 classes?

Thanks!

This is a deprecated feature, but I still need to implement it for compatible.

We have faced with the problem of extending server side api.
We added optional field at the end of the message:
optional uint64 creation_date = 11;

Flash client fails with next error(java and c clients work correctly):
Exception fault: Error: Invalid wire type: 3
 at com.netease.protobuf::ReadUtils$/skip()[.../ReadUtils.as:32]

bytes for this fields look like - 58a080a3eb8227
58 - field descriptor
a080a3eb8227 - data

According to code of ReadUtils:
case WireType.VARINT:
   while (input.readUnsignedByte() > 0x80) {}

but we have second byte = 80 so the parsing is stopped
and next bytes are not read correctly

In certain cases a null object reference error is thrown when trying to run 
toString on a protobuf object.

TypeError: Error #1009: Cannot access a property or method of a null object 
reference.
at 
-0Rv::TextFormat$/-0I2()C:\Users\Atry\Desktop\protoc-gen-as3\as3\com\netease\pro
tobuf\TextFormat.as:290
at 
-0Rv::TextFormat$/-I8()C:\Users\Atry\Desktop\protoc-gen-as3\as3\com\netease\prot
obuf\TextFormat.as:199
at 
-0Rv::TextFormat$/-0Cm()C:\Users\Atry\Desktop\protoc-gen-as3\as3\com\netease\pro
tobuf\TextFormat.as:309
at 
-0Rv::TextFormat$/-0PH()C:\Users\Atry\Desktop\protoc-gen-as3\as3\com\netease\pro
tobuf\TextFormat.as:319
at 
_-0Rv::Message/toString()C:\Users\Atry\Desktop\protoc-gen-as3\as3\com\netease\pr
otobuf\Message.as:209

如题

See 
https://code.google.com/apis/protocolbuffers/docs/reference/cpp/google.protobuf.
text_format.html#TextFormat.Printer.SetUseShortRepeatedPrimitives.details

http://code.google.com/intl/zh-CN/apis/protocolbuffers/docs/proto.html#services

Lets say you have a protobuf:

message Person {
    optional string name = 1;
}

Inside the Person AS3 file you'll have:

private var name_:int;
private var hasName_:Boolean = false;
public function get hasName():Boolean {
    return hasName_;
}
public function set name(value:String):void {
    hasName_ = true;
    name_ = value;
}
public function get name():String {
    return name_;
}

My proposition is to add another method:

public function removeName():void {
    hasName_ = false;
    name_ = null;
}

Which would be handled like expected in writeToBuffer():

if (hasName) {
    WriteUtils.writeTag(output, WireType.LENGTH_DELIMITED, 1);
    WriteUtils.write_TYPE_STRING(output, name);
}

Doing 'person.name = null' simply sets name to null. It doesn't set hasName to 
false.

协议内容如下：

message C10001{
    required string account = 1;
    required bytes password = 2;
}

用 as3crypto 的 sha256 对 password 
进行处理后进行字用protoc-gen-as3串化，打印出来发现跟C++生成
的二进制数据有一位不一样，目前分析Sha256后的数据两种语��
�是一直的，说以不知道问题是不是出现在串化过程中
两种语言串化的二进制数据如下（不包含括号，括号是我加��
�去突出不同的那个位置的）
AS3:6433130303031a561646d696e1220ba7816bf8f1cfea414140de5dae2223b0361a396177a9cb
410ff61f2
C++:6433130303031a561646d696e121dba7816bf8f1cfea414140de5dae2223b0361a396177a9cb
410ff61f2

AS的代码如下：
import com.hurlant.crypto.Crypto;
            import com.hurlant.crypto.hash.IHash;
            import com.hurlant.util.Hex;
            import com.netease.protobuf.*;

            import msg.*;

            private function CK():void{

                var hash:IHash = Crypto.getHash("sha256");
                var data1:ByteArray = Hex.toArray(Hex.fromString("abc"));

                var msgA:C10001 = new C10001();
                msgA.account = "admin";
                msgA.password = hash.hash(data1);

                var cN:String = getQualifiedClassName(msgA);
                cN=cN.substr(cN.lastIndexOf(':')+1);

                var data:ByteArray = new ByteArray();
                data.writeByte(cN.length);
                data.writeUTFBytes(cN);

                msgA.writeTo(data);

                var code:String = "";
                for(var i:int=0; i < data.length; i++){
                    code += data[i].toString(16);
                }

                trace(code);

            } 

What steps will reproduce the problem?
1. Run protoc with the protoc-gen-as3 plug-in
2. Compile the generated code with the Flex 4 SDK compiler
3. Observe the warning described below

What is the expected output? What do you see instead?

Here is an example of the message:
[WARNING] C:\...\Projects\foo\src\main\flex\protocol\MapInfoMessage.as:[10,21] 
No constructor function was specified for class MapInfoMessage.        public 
final class MapInfoMessage extends com.netease.protobuf.Message implements 
flash.utils.IExternalizable {


What version of the product are you using? On what operating system?

0.8.8

Please provide any additional information below.

Currently optinal or required fields may be watched using data bindings. This 
does not work for repeated (array) fields, since arrays send no events.

A solution for this would be a field option "as3_array_collection":

repeated int32 foo = 1 [(as3_array_collection) = true];

This option changes the type of the repeated field to ArrayCollection.

This enables the use of the field in a data binding as source for e.g. a spark 
list, that will update as soon as the field on the message is changed.

您好，现在我们有个新项目客户端打算用flash来做，服务器用
java。消息协议我打算用protobuf。
，我想问一下就是我之前java端protobuff版本是2.5.x，看了下proto
buff-gen-as用的是2.4.1，请问这个有没有影响，需不需要版本统�
��？

比如说
message Aaa {
  message User {}
}
message Bbb {
  message User {}
}
就会生成两个User类在两个不同的package下，但是导入使用就很
麻烦
或许生成AaaUser和BbbUser类比较好

What steps will reproduce the problem?
1. Create a .proto file with a message containing a required field
2. Compile the proto to an as file
3. Call toString() against the generated message

What is the expected output? What do you see instead?
field name and value of the required field.  Get an empty string instead

What version of the product are you using? On what operating system?
1.0.0rc3 on Windows 7 64bit

Please provide any additional information below.

Branch name:
mail-as3-service
http://code.google.com/r/mail-as3-service/source/list

Purpose of code changes on this branch:
# 生成 as3 的 service interface，方便实现从另一端调用 as3 
提供的服务，另外已经据此实现了 rpc，见 
http://code.google.com/p/abu-rpc/source/browse/#svn%2Ftrunk%2Fsrc%2Factionscript
3

# 生成的代码见 
http://code.google.com/p/abu-rpc/source/browse/trunk/src/actionscript3/echo/src/
EchoService%24Interface.as 已经按之前的建议提供 done 
函数，提供异步返回结果的机制

# （与原实现不兼容之处）修改了 stub 类的命名，全部加了 
$Stub 后缀，例见 
http://code.google.com/p/abu-rpc/source/browse/trunk/src/actionscript3/echo/src/
EchoService%24Stub.as

# 按上次修改意见，已经修改了 writeIService 之类的函数。

When reviewing my code changes, please focus on:


After the review, I'll merge this branch into:
/trunk

Thanks a lot for the library. It is very useful.

com.netease.protobuf.Message class changes the input's endian attribute when 
reading data.

I think it should be restored. I have had problems reading from a socket. I was 
reading data encoding with protobuf but also unencoded data. The first 
socket.readShort() was correct but the next, after reading a protobuf chunk,  
were wrong. I solved it setting back the endian attr to 
flash.utils.Endian.BIG_ENDIAN but I'd prefer that the library didn't interfere.

ACTUAL CODE:
public final function mergeFrom(input:IDataInput):void {
    input.endian = flash.utils.Endian.LITTLE_ENDIAN
    readFromSlice(input, 0)
}
public final function mergeDelimitedFrom(input:IDataInput):void {
    input.endian = flash.utils.Endian.LITTLE_ENDIAN
    ReadUtils.read$TYPE_MESSAGE(input, this)
}

PROPOSED CODE:
public final function mergeFrom(input:IDataInput):void {
    const endian:String = input.endian
    input.endian = flash.utils.Endian.LITTLE_ENDIAN
    readFromSlice(input, 0)
    input.endian = endian
}
public final function mergeDelimitedFrom(input:IDataInput):void {
    const endian:String = input.endian
    input.endian = flash.utils.Endian.LITTLE_ENDIAN
    ReadUtils.read$TYPE_MESSAGE(input, this)
    input.endian = endian
}

"Packed" is a option for repeated fields.

Translate wiki pages to English

生成的as3代码中在writeToBuffer
里面经常会看到类似的情况：
for (var heros$index:uint = 0; heros$index < this.heros.length; ++heros$index) {

－》》

for (var heros$index:uint = 0,len:uint= this.heros.length; heros$index <len; 
++heros$index) {

虽然不是很关键，但是看着总是觉得应该优化一下，希望你��
�时间时参考以下，谢谢⋯⋯

What steps will reproduce the problem?
1. Have a protocol with a message containing one int32
2. Send one such message with a negative value for the int, for instance 
the Java implementation
3. Try to decode it in AS3

What is the expected output? What do you see instead?

The message should be properly decoded (int32 is allowed to take negative 
values) but crashes the lib instead. This seems to come from the assumption 
that int32 are positive made by the lib. I'm looking into fixing it, but 
feedback would be welcome.


Please provide any additional information below.

What steps will reproduce the problem?

Run protoc with the protoc-gen-as3 plugin, specifying the attached .proto 
file

What is the expected output? What do you see instead?

On compiling the generated code (flex 3.2), observe the following errors:
[ERROR]...protocol\ChatMessage.as:[10,41] Base class is final.        
public final class ChatMessage extends Message implements IExternalizable {
[ERROR]...protocol\ChatMessage.as:[26,31] Method marked override must 
override another method.                protected override function 
writePostposeLength(output:PostposeLengthBuffer):void {

What version of the product are you using? On what operating system?

Version 0.8.6 of protoc-gen-as3
Version 3.2 of the Flex SDK
Windows

Please provide any additional information below.

use the newest, output parse error,

back 1.1.0, ok!

the build is currently broken:

make: *** No rule to make target `tiny-protobuf.swc', needed by 
`dist/protobuf.swc'.  Stop.

Branch name:

Purpose of code changes on this branch:


When reviewing my code changes, please focus on:


After the review, I'll merge this branch into:
/trunk

What steps will reproduce the problem?
1. Sender sends two payloads in quick succession
2. Receiver is unable to read entire payload off of socket before second 
payload hits.
3. Receiver begins reading the second payload as part of the first payload, 
adding payload 2's data to payload 1 and/or throwing an error when payload 2 
has a common field with payload 1 and it tries to assign payload 2's value to 
the already assigned field from payload 1

What is the expected output? What do you see instead?
The expected output would be the correct reception of two unique payloads.
Instead you see a single payload consisting of 2 payloads data or an error when 
reading the second payload.

What version of the product are you using? On what operating system?
3.0.8.9

Please provide any additional information below.

将字符串格式化为字节数组时，能不能指定编码为unicode呀？�
��不单是utf-8，希望
改进，谢谢。

What steps will reproduce the problem?
1. create a proto with a field called 'tag' and type string
2. run protoc 
3. try to compile the code 

What is the expected output? What do you see instead?
Should compile, instead complains that tag is a string and cannot assign int to 
string (local variable tag conflicts with class variable 'tag' from the proto)

What version of the product are you using? On what operating system?
0.9.1 - but i see the code is unchanged in 1.0 RC.

Please provide any additional information below.
Please simply rename the local variable in the readFromSlice() method to 
something unique to prevent these conflicts. 

What steps will reproduce the problem?
1. download and build the 2.3.1 repo of protoc
2. download and extract protoc-gen-as3-0.8rc2-bin.tar.gz
3. run protoc with the following command line:
/home/orion/sandbox/orion/pb2.3.0/linux/bin/protoc --plugin=protoc-gen-
as3=/home/orion/sandbox/orion/pb2.3.0/protoc-gen-as3/dist/protoc-gen-as3 
--as3_out foo ./xs1.proto
(full paths included in case i'm doing something dumb)

What is the expected output? What do you see instead?
the result is the following error:
-----------------
/home/orion/sandbox/orion/pb2.3.0/protoc-gen-as3/dist/protoc-gen-as3: program 
not found or is not executable
--as3_out: protoc-gen-as3: Plugin failed with status code 1.
-----------------

if i remove "-e" from the first line of protoc-gen-as3,
then the error changes to:
-----------------
Exception in thread "main" java.lang.NoClassDefFoundError: 
com/netease/protocGenAs3/Main
Caused by: java.lang.ClassNotFoundException: com.netease.protocGenAs3.Main
    at java.net.URLClassLoader$1.run(URLClassLoader.java:200)
    at java.security.AccessController.doPrivileged(Native Method)
    at java.net.URLClassLoader.findClass(URLClassLoader.java:188)
    at java.lang.ClassLoader.loadClass(ClassLoader.java:303)
    at sun.misc.Launcher$AppClassLoader.loadClass(Launcher.java:301)
    at java.lang.ClassLoader.loadClass(ClassLoader.java:248)
    at java.lang.ClassLoader.loadClassInternal(ClassLoader.java:316)
Could not find the main class: com/netease/protocGenAs3/Main. Program will exit.
--as3_out: protoc-gen-as3: Plugin failed with status code 1.
-----------------


What version of the product are you using? On what operating system?

these results are consistent on OS X (snow leopard) and on a centos linux box:
Linux version 2.6.18-164.6.1.el5 ([email protected])



Please provide any additional information below.

在学习您的类库的时候发现传给服务器的浮点数 和 
接收到的浮点数最后几位不一样。100%出现。

What steps will reproduce the problem?

1. Create [.proto] file.

    message DoubleFloat {
       optional double myDouble = 1;
       optional float myFloat = 2;
    }

2. Run this program.

    var values:DoubleFloat = new DoubleFloat();
    values.myDouble = 1.0;
    values.myFloat  = 1.0;

    var bytes:ByteArray = new ByteArray();
    values.writeTo(bytes);

    trace(bytes[0].toString(16));   // 09 -- correct
    trace(bytes[1].toString(16));   // 00 -- correct
    trace(bytes[2].toString(16));   // 00 -- correct
    trace(bytes[3].toString(16));   // 00 -- correct
    trace(bytes[4].toString(16));   // 00 -- correct
    trace(bytes[5].toString(16));   // 00 -- correct
    trace(bytes[6].toString(16));   // 00 -- correct
    trace(bytes[7].toString(16));   // F0 -- correct
    trace(bytes[8].toString(16));   // 3F -- correct

    trace(bytes[9].toString(16));   // 15 -- correct
    trace(bytes[10].toString(16));  // 00 -- correct
    trace(bytes[11].toString(16));  // 00 -- correct
    trace(bytes[12].toString(16));  // 80 -- correct
    trace(bytes[13].toString(16));  // 3F -- correct

    bytes.position = 0;
    var loaded:DoubleFloat = new DoubleFloat();
    loaded.readFromSlice(bytes, 0);

    bytes.clear()
    loaded.writeTo(bytes);

    trace(bytes[0].toString(16));   // 09 -- correct
    trace(bytes[1].toString(16));   // 3F -- wrong(It is big endian!)
    trace(bytes[2].toString(16));   // F0 -- wrong(It is big endian!)
    trace(bytes[3].toString(16));   // 00 -- correct
    trace(bytes[4].toString(16));   // 00 -- correct
    trace(bytes[5].toString(16));   // 00 -- correct
    trace(bytes[6].toString(16));   // 00 -- correct
    trace(bytes[7].toString(16));   // 00 -- wrong(It is big endian!)
    trace(bytes[8].toString(16));   // 00 -- wrong(It is big endian!)

    trace(bytes[9].toString(16));   // 15 -- correct
    trace(bytes[10].toString(16));  // 3F -- wrong(It is big endian!)
    trace(bytes[11].toString(16));  // 80 -- wrong(It is big endian!)
    trace(bytes[12].toString(16));  // 00 -- wrong(It is big endian!)
    trace(bytes[13].toString(16));  // 00 -- wrong(It is big endian!)

    bytes.position = 0;
    var loaded2:DoubleFloat = new DoubleFloat();
    loaded2.readFromSlice(bytes, 0);

    bytes.clear()
    loaded2.writeTo(bytes);

    trace(bytes[0].toString(16));   // 09 -- correct
    trace(bytes[1].toString(16));   // 00 -- correct
    trace(bytes[2].toString(16));   // 00 -- correct
    trace(bytes[3].toString(16));   // 00 -- correct
    trace(bytes[4].toString(16));   // 00 -- correct
    trace(bytes[5].toString(16));   // 00 -- correct
    trace(bytes[6].toString(16));   // 00 -- correct
    trace(bytes[7].toString(16));   // F0 -- correct
    trace(bytes[8].toString(16));   // 3F -- correct

    trace(bytes[9].toString(16));   // 15 -- correct
    trace(bytes[10].toString(16));  // 00 -- correct
    trace(bytes[11].toString(16));  // 00 -- correct
    trace(bytes[12].toString(16));  // 80 -- correct
    trace(bytes[13].toString(16));  // 3F -- correct

What is the expected output? What do you see instead?

    The ByteArray should keep same values anytime.

What version of the product are you using? On what operating system?

    1.0.0-rc6

Please provide any additional information below.

The Makefile currently targets Flash Player 10, which I don't think is 
available/supported anymore (my copy of the Flex SDK 4.6 doesn't contain 
support for it).

I updated my Makefile to target Flash Player 11 to get things compiling again

Hello,

I'm not reporting a bug issue here. I'm just curious about two compilation 
warnings reported by Flash Builder 4.6. I get those warnings on the as3 classes 
I generated through protoc-gen-as3. However my Flash Builder project compiles & 
runs perfectly fine (serialization & deserialization works).

Here are the two warnings (see attached RequestHeader.as file):
* first one is on the following line:
        import com.netease.protobuf.fieldDescriptors.*;
  it says "the import fieldDescriptors could not be found"
* second one is on the following line:
        public static const REQUESTID:FieldDescriptor$TYPE_INT32 = new FieldDescriptor$TYPE_INT32
  it says "type was not found or was not a compile-time constant: FieldDescriptor$TYPE_INT32"

My Question:
com.netease.protobuf.fieldDescriptors package doesn't exist inside 
protobuf.swc. Could you please explain to me what are those both warnings ? Are 
they perfectly normals ?

What steps will reproduce the problem?
1. generate as3 classes from the attached sample.proto file
2. Create a new Flash Builder project (AIR project in my case)
3. Add the as3 classes generated from sample.proto to the newly created Flash 
Builder project
4. Add the protoc-gen-as3 library (protobuf.swc) to the Referenced Libraries of 
the Flash Builder project

What version of the product are you using? On what operating system?
protoc-gen-as3 3.1.0.0-rc9
Flash Builder 4.6
Windows Vista Integral Edition SP2 x64

Best Regards,
--Seb

我现在每个包都是一个message.
然后每个message都是由head和type开头的,tag也都一样,只是默认值
不一样.
optional int32 head = 1;
optional int32 type = 2;
比如message LoginOKResponder的定义如下:
message LoginOKResponder {
optional int32 head = 1 [default = 1];
optional int32 type = 2 [default = 1];
required string id = 3;
required string name = 4;
}

在java中使用CodedInputStream可以直接readInt32得到head和type,
CodedInputStream input = CodedInputStream.newInstance(data);
int header = input.readInt32();
int action = input.readInt32();
这样会得到header 和 action.

但是在AS中我找不到类似的类或方法, 
用ReadUtil读出来的是错误的值.
然后我定义了SimpleResponder,结构如下:
message SimpleResponder {
optional int32 head = 1;
optional int32 type = 2;
}
我本来想让SimpleResponder 
帮我把head和type读出来,但是发现读出来的也是错的.

希望作者能帮忙实现我这样的需求.
我的每个message都有head和type,然后想在不知道具体的message类的
情况下读到head和type.

当然也可以把head和type从message里分出来.但是我想使用proto的��
�认值,然后就不用在代码里写了,而且用message包起来也方便传�
��的.

最后,谢谢了.希望作者能实现我这个需求.

Add a flag indicate generated code should use haxe.Int64 instead of 
com.netease.protobuf.Int64 and com.netease.protobuf.Int64 .

In com.netease.protocGenAs3.Main.java, there are a couple AssertionErrors that 
are thrown with both a String and an Exception. The AssertionError(String, 
Throwable) constructor is not available in Java 6 and below.

Removing the exception argument will allow compilation on Java6 and earlier.

What steps will reproduce the problem?
1. Adobe Air SDK package ipa
2. SDK15 without -useLegacyAOT yes
3. SDK16 

What is the expected output? What do you see instead?
Compilation failed while executing : compile-abc

Please provide any additional information below.
TextFormat.as
The functoin skipWhitespace() has an infinite loop.
You can fix this the bug by Using:
private static function skipWhitespace(source:ISource):void {
        for (;; ) {
                const b:int = source.read()
                switch (b) {
                case 0x20:/* space */
                case 0x09:/* \t */
                case 0x0a:/* \n */
                case 0x0d:/* \r */
                        continue
                case 0x23:/* # */
                        for (;;) {
                                var isLineEnd:Boolean = false;
                                switch (source.read()) {
                                        case 0x0a:/* \n */
                                        case 0x0d:/* \r */
                                                isLineEnd = true;
                                                break;
                                        default:
                                                break;
                                }
                                if(isLineEnd)
                                {
                                        break;
                                }
                        }
                        break;
                default:
                        source.unread(b)
                        return
                }
        }
}

禁止在 LENGTH_DELIMITED 字段中使用 packed = true
收包时判断一个 repeat 字段是否 packed 要根据 wire type 
而不是编译时选项。

目前看到的 rpc 
实现基本上都传递方法名给另一端进行函数查找，当跨语言��
�时候，as3 发过去的是首字母小写的骆驼风格，C++/Python 
的则是没有经过处理的，所以无法查找到相应的方法。
建议不要修改 rpc 
的函数名，业务开发人员怎么写的，就怎么用，类似 message 
名的处理方法。在这方面，protobuf 的 style guide 也提出 rpc 
名应统一使用首字母大写的驼峰法。见 
https://developers.google.com/protocol-buffers/docs/style?hl=zh-CN

1. Defining a nested message as follow:

message Test{
    message Nested{
        required int32 id = 1;
    }
    repeated Nested nested = 1;
}
2. Generated code as follow

public final function readFromSlice(input:flash.utils.IDataInput, 
bytesAfterSlice:uint):void {
    ...
    case 1:
           const nested$element:MessageDefinition.Test.Nested = new MessageDefinition.Test.Nested();
    ...

3. 'const' is causing compiling error

What steps will reproduce the problem?
1. client & server have the same protocol.
2. 1 new token added to enum. server updated to use new protocol.
3. server send message to the client(with old api) with value set to new token 
from enum.
4. IOError(value + " is invalid for " + enumTypeDescription.@name) thrown on 
message.toString()

What is the expected output? What do you see instead?
no Error thrown

What version of the product are you using? On what operating system?
2.4.1


Please provide any additional information below.

What steps will reproduce the problem?
1. download and build the 2.3.1 repo of protoc
2. download and extract protoc-gen-as3-0.8rc2-bin.tar.gz
3. run protoc with the following command line:
/home/orion/sandbox/orion/pb2.3.0/linux/bin/protoc --plugin=protoc-gen-
as3=/home/orion/sandbox/orion/pb2.3.0/protoc-gen-as3/dist/protoc-gen-as3 
--as3_out 
foo ./xs1.proto
(full paths included in case i'm doing something dumb)

What is the expected output? What do you see instead?
the result is the following error:
-----------------
/home/orion/sandbox/orion/pb2.3.0/protoc-gen-as3/dist/protoc-gen-as3: program 
not 
found or is not executable
--as3_out: protoc-gen-as3: Plugin failed with status code 1.
-----------------

if i remove "-e" from the first line of protoc-gen-as3,
then the error changes to:
-----------------
Exception in thread "main" java.lang.NoClassDefFoundError: 
com/netease/protocGenAs3/Main
Caused by: java.lang.ClassNotFoundException: com.netease.protocGenAs3.Main
    at java.net.URLClassLoader$1.run(URLClassLoader.java:200)
    at java.security.AccessController.doPrivileged(Native Method)
    at java.net.URLClassLoader.findClass(URLClassLoader.java:188)
    at java.lang.ClassLoader.loadClass(ClassLoader.java:303)
    at sun.misc.Launcher$AppClassLoader.loadClass(Launcher.java:301)
    at java.lang.ClassLoader.loadClass(ClassLoader.java:248)
    at java.lang.ClassLoader.loadClassInternal(ClassLoader.java:316)
Could not find the main class: com/netease/protocGenAs3/Main. Program will exit.
--as3_out: protoc-gen-as3: Plugin failed with status code 1.
-----------------


What version of the product are you using? On what operating system?

these results are consistent on OS X (snow leopard) and on a centos linux box:
Linux version 2.6.18-164.6.1.el5 ([email protected])



Please provide any additional information below.

I would like to have a feature like 'option java_package'.  For example, if I 
added:

option as3_package = "com.example.foo"

Then the generated classes would end up in the package com.example.foo.

C:\Users\admin\Desktop\protoc-gen-as3-1.1.0-bin>protoc.exe 
--plugin=protoc-gen-as3=C:\Users\admin\Desktop\protoc-gen-as3-1.1.0-bin\protoc-g
en-as3.bat -as3_out=C:\Users\admin\Desktop\protoc-gen-as3-1.1.0-bin 
C:\Users\admin\Desktop\protoc-gen-as3-1.1.0-bin\a.proto
Unknown flag: -a

在命令行里执行，出现未知的标记 -a 不知道是怎么回事。

protoc.exe用2.3 和 2.4.1都出现这个问题。

What steps will reproduce the problem?
1. Send float from as3 to c
2. The variable Number is 123 (example)
3.

What is the expected output? What do you see instead?
A float. it arrives a 0

What version of the product are you using? On what operating system?
last

Please provide any additional information below.

If we want a parse the data like 123.12 is working ok, but if the number is an 
integer or the are more numbers after the comma, it goes crazy.

message ParentMessage {
    extensions 1 to max;
}

extend ParentMessage {
    optional ExtensionMessage extension_message = 1;
}

message ExtensionMessage {
}

This will generate a file EXTENSION_MESSAGE.as that initializes 
ParentMessage.extensionReadFunctions.

This will generate a file initializer.as.inc that references EXTENSION_MESSAGE.

The Bug: If you include these files in a program in an ActionScript project in 
Flash CS4, the references to EXTENSION_MESSAGE are ignored.  
ParentMessage.extensionReadFunctions is not initialized.