Skip to main content
Version: Next

Code and Data Generation

Support mainstream game engines and platforms​

  • unity
  • unreal
  • cocos2d-x
  • godot
  • egret
  • WeChat Mini Game Platform
  • Other small game platforms that support js

Support mainstream thermal solutions​

  • HybridCLR
  • {x,t,s...}lua
  • ILRuntime
  • puerts
  • other
  • skynet
  • ET
  • GameFramework
  • QFramework
  • other

Support mainstream game development languages​

  • c# (.net framework 4+. dotnet core 3+)
  • java (1.6+)
  • go (1.10+)
  • lua (5.1+)
  • typescript (3.0+)
  • python (3.0+)
  • c++ (11+)。
  • erlang (18+). The classic version is supported by Luban and has not been migrated to the current version yet.
  • rust (1.5+). The classic version is supported by Luban and has not been migrated to the current version yet.
  • godot. The classic version is supported by Luban and has not been migrated to the current version yet.
  • php
  • dart

It is very easy to customize or extend support for new languages.

Supported Data Formats​

tip

In the same format, the data generated for different languages is exactly the same

  • binary format. The format is compact and efficient to load, but basically unreadable. Recommended for official releases only.
  • bin-offset format. Record the index position of each record in the data file exported in bin format, which can be used for lazy loading with record granularity|
  • standard json format
  • protobuf bin and json
  • flatbuffers json
  • msgpack
  • -lua
  • -xml
  • -erlang
  • -yaml

bin-offset format​

Sometimes you don't want to load the entire table directly, but you want to use the record as the granularity, and then load a certain record when you access the record. The bin-offset format records the offset of each record in the bin file in the bin format, so that when a record is accessed, if it is not loaded, the data at the corresponding offset is directly read from the bin file.

bin-offset serializes the key and offset information of each record in record order.

| record_index 1 | record_index 2| ... | record_index n|

The implementation of record_index k is to serialize all the keys of the record, and then serialize the offset of the record.

buf. Write(key1);
buf. Write(key2);
...
buf. Write(Key N);
buf. WriteSize(offset);

It can be understood more clearly directly from the source code of bin-offset.

     // x is the output bin-offset file
     private void WriteList(DefTable table, List<Record> datas, ByteBuf x)
     {
         // buf corresponds to the output bin file
         ByteBuf buf = new ByteBuf();
         buf. WriteSize(datas. Count);
         foreach (var d in datas)
         {
             foreach (var indexInfo in table. IndexList)
             {
                 DType keyData = d.Data.Fields[indexInfo.IndexFieldIdIndex];
                 // Serialize each key of the record
                 keyData.Apply(BinaryDataVisitor.Ins, x);
             }
             // The offset of the serialized record
             x. WriteSize(buf. Size);
             d.Data.Apply(BinaryDataVisitor.Ins, buf);
         }
     }

The formats supported by different languages are as follows:​

In the same language, different codes need to be generated for loading different data formats. That is, the code target and data target must match.

languagebinaryjsonlua
c#✔️✔️
java✔️✔️
go✔️✔️
lua✔️✔️
c++✔️
go✔️✔️
python✔️
typescript✔️✔️
php✔️
rust✔️
godot✔️
dart✔️
protobuf✔️✔️

generate​

For details, see Command-line Tools.