a JS bytecode compiler for Yarn
npm install jacquard-yarncompilerThis compiles Yarn syntax (with help of the yarn parser into a simpler bytecode suitable for running in games/browsers
Install jacquard-yarncompiler via your favourite javascript package tool.
API Docs are here
Current version is 0.1.0
All groupings are Little Endian. All Strings are UTF-8.
#### Fixed Length
##### Float
1. 4 bytes - the IEEE 754 float value
##### Byte
1. 1 byte - usually either a bit-field or a unsigned integer.
##### StackOffset
1. 1 byte - the offset (0 being the top of the stack and 255 being the 255th oldest item on the stack) of a stack argument
#### Variable Length
##### VarInt
1. Type (Byte) - length + signed or unsigned - high bit is on if it's signed, the lower 7 are the NumBytes of the number (current max length is 8 bytes). If this is 0 then there's no bytes and the value is 0.
2. then NumBytes of
1. NumberPart (Byte) - Part of the number
##### VarBytes
1. NumBytes (VarInt) - indicating how many bytes are in this value.
2. then NumBytes of
1. Part (Bytes) - part of this value.
##### VarString
A Varbytes where the Bytes value represents an UTF-8 encoded string value.
##### StringTable
1. NumEntries (VarInt) - how many entries does this table have.
2. and then NumEntries of
1. Entry (VarString) - the string at this entry.
##### BytesEntryPointTable
1. NumEntries (VarInt) - how many entries does this table have.
2. and then NumEntries of
1. Entry (VarBytes) - the byte name of this entry.
2. Offset (VarInt) - the byte offset for this entry in the instruction set.
##### StringEntryPointTable
1. NumEntries (VarInt) - how many entries does this table have.
2. and then NumEntries of
1. Entry (VarString) - the string at this entry.
2. Offset (VarInt) - the byte offset for this entry in the instruction set.
The logic file contains the general graph flow and expressions. Text that is to be displayed on screen is not in this file as that needs localization.
- ASCII "JQRDL" - denoting file type (should take 5 bytes)
- Version (VarString) - the bytecode version.
- Function Table Index (VarInt) - Offset (in entire file) of where the function table starts (or 0 if it doesn't exist)
- Character Table Index (VarInt) - Offset (in entire file) of where the character table starts (or 0 if it doesn't exist)
- Variable Table Index (VarInt) - Offset (in entire file) of where the variable table starts (or 0 if it doesn't exist)
- String Table Index (VarInt) - Offset (in entire file) of where the string table starts (or 0 if it doesn't exist)
- Node Table Index (VarInt) - Offset (in entire file) of where the Node table starts
- Logic Instruction Block Index (VarInt) - Offset (in entire file) of where the Instruction table starts
- Function Table (StringTable) - The Function table (if exists)
- Character Table (StringTable) - The Character table (if exists)
- Variable Table (StringTable) - The Variable table (if exists)
- String Table (StringTable) - The String table (if exists)
- Node Table (StringEntryPointTable) - The Node Table (with entry points)
- Logic Instruction Block - Defined below
#### Logic Instruction Block
This is a large bucket of bytes that is back-to-back opcodes and opcode arguments and defines the actual program flow. Use the Node Table (and a node name) to get your entry point (or alternatley, start at the first instruction). Consume the opcode you are at and proceed accordingly.
There is one (or more) of these files - each one representing one language. This file should no branching logic or expressions that aren't involved in text display on screen / for subtitling or VO purposes.
- ASCII "JQRDD" - denoting file type (should take 5 bytes)
- Version (VarString) - the bytecode version
- Language (VarString) - The identifier for this language (be sure to be consistent in your project)
- Character Table Index (VarInt) - Offset (in entire file) of where the character table starts (or 0 if it doesn't exist)
- String Table Index (VarInt) - Offset (in entire file) of where the string table starts (or 0 if it doesn't exist)
- Dialog Instruction Block Index (VarInt) - Offset (in entire file) of where the Dialog Block table starts
- Character Table (StringTable) - The Character table (if exists) - these characters are localized for this language and are in the same order as the logic file
- String Table (StringTable) - The String table (if exists)
- Dialog Block Table (BytesEntryPontTable) - The table of dialog blocks for this language.
- Dialog Instruction Block - Defined below
#### Dialog Instruction Block
This is a large bucket of bytes that is back-to-back opcodes and opcode arguments and defines text display for a given languagae. Use the Dialog Block table to get your entry point and run until you encounter the DialogBlockEnd opcode.
Documented here
Opcodes are variable length - the opcode itself is always one byte, but the arguments may be more. Types are listed above
Jacquard - Yarn Parser requires the following dependencies be pre-installed:
* NodeJS (https://nodejs.org).
* Yarn package manager (https://yarnpkg.com).
First, you need to make sure that you have nodeJs and Yarn installed (see the dependencies section above and install the relevant version for your platform).
Currenlty we're using node v8.9.4 and yarn v1.5.1.
Using your command line, navigate to the directory you've cloned this repo into.
1. Execute a yarn install. This will download all the necessary libraries you need.
2. You can then do a compile via a yarn compile specifying the source yarn files (or a directory containing multiple files) and the destination directory for the compiled bytecode and other files
If you wish to rebuild the documentation, you'll need to have documentation.js installed, which you can do via npm install -g documentation - then you can run yarn serveDoc to serve the docs locally or yarn buildDoc to build the docs directory