Command Block Assembly
Command Block Assembly started off as a tool that compiled assembly instructions into Minecraft commands.
It has now grown into a much larger suite of tools for compiling source code into Minecraft commands, therefore the toolchain as a whole is known as the Minecraft Compiler Collection (MCC).
Minecraft Compiler Collection
The Minecraft Compiler Collection (MCC) is an umbrella tool composed of multiple compilers in a toolchain to compile high level code into Minecraft commands. It shares some similarities with the GNU Compiler Collection.
MCC is composed of the following components:
CBL Compiler
Command Block Language (CBL) is a C++ inspired programming language specifically designed for Minecraft commands.
CBL has been designed to abstract away from commands to the point where you don't need to consider the underlying mechanics.
The syntax is similar to C++ in a lot of ways. It also takes from Java's idea of having no pointers in the language.
Example
include "Text"
type point {
int x;
int y;
constructor(int x, int y) {
this.x = x;
this.y = y;
}
inline point operator +(point other) {
return point(this.x + other.x, this.y + other.y);
}
inline point operator *(int val) {
return point(this.x * val, this.y * val);
}
inline void print();
}
inline void point::print() {
Text output;
output += "(";
output.append_ref(this.x);
output += ", ";
output.append_ref(this.y);
output += ")";
output.send_to_all();
}
void main() {
point p1(2, 4);
point p2(5, 9);
p1 += p2;
p1 *= 2;
p1.print();
}
This example demonstrates several language features of CBL, some of which should be familiar to C++ programmers.
Because everything in this example can be determined statically (at compile time), the output is just:
tellraw @a [{"text":"("},{"text":"14"},{"text":", "},{"text":"26"},{"text":")"}]
Documentation
See the CBL Wiki for more details on the CBL language.
Assembler
The assembler in MCC is the original assembler when this project was just Command Block Assembly.
Assembly Language
The assembly language is simple and has instructions similar to x86.
The language looks like the following:
#include foo.asm
; Useful comment
.my_const #1 ; A constant value
.my_ref my_const ; A constant reference
main:
MOV #0x01, 0 ; Do the thing
_loop: ADD my_const, 0
JMP _loop
Example
The first example code written for CBA was an assembly program that prints the fibinacci sequence until the next value overflows:
.x 0x00
.y 0x01
.old_x 0x02
.counter 0x03
main:
MOV #0, x
MOV #1, y
MOV #1, counter
_loop:
PRINT "fib(", counter, ") = ", x
SYNC
ADD #1, counter
MOV x, old_x
MOV y, x
ADD old_x, y
CMP #0, x
JGE _loop ; if not >=0 then x has overflowed
Documentation
The instruction set and how write CBA programs is documented on the wiki.
C Compiler
MCC partially implements a C compiler, most language features are implemented and it includes a preprocessor.
The C compiler sits on top of the assembler - all C code is compiled into assembly which is then passed through MCC's assembler.
Example
Here is the same fibinacci sequence program but implemented in C:
#include <stdio.h>
int x;
int y;
int old_x;
int counter;
void main() {
x = 0;
y = 1;
counter = 1;
do {
printf("fib(%d) = %d", counter++, x);
sync;
old_x = x;
x = y;
y += old_x;
} while(x >= 0);
}
Documentation
The C compiler tries to stay as close as possible to the real C language, so documentation for most language features can be found elsewhere.
There is only one built-in type, int
. Fundamentally, all data is stored in a scoreboard objective
which is a 32 bit signed integer.
The mclib.h file contains several useful macros and definitions.
Browse the code in the examples directory to see working examples of C code.
Command IR Compiler
Command IR is the intermediate representation used by MCC. All compilers above ultimately generate Command IR code.
It is designed to relate closely to Minecraft commands themselves but provide a level of abstraction and context which enables optimizations among other things.
Command IR supports linking, which means it's possible to write code in CBL, C, ASM and IR directly and compile them together into one program.
Example
Here is a hello world program:
function hello {
preamble {
$all_players = selector a
$message = text
extern
}
compiletime {
entry:
text_append $message, "Hello, "
text_append $message, "World!"
}
begin:
text_send $message, $all_players
ret
}
The output is a file hello.mcfunction
:
tellraw @a [{"text":"Hello, "},{"text":"World!"}]
Documentation
Command IR's syntax, instruction reference and internal working is documented on the wiki.
Datapack Packer
Finally, to bring everything together and generate a datapack, MCC reads a "Datapack Definition" file which declares how to package the code into a datapack.
The file is a simple INI file declaring attributes such as the namespace.
To complete the fibinacci example, the datapack definition file fib.dpd
is:
[Datapack]
namespace=fib
place location=0 56 0
description = An example datapack that prints the fibonacci sequence
Documentation
Currently there is only one section in a DPD file, the Datapack section.
In order to create a datapack, two values are required: the namespace
and the place location
Option | Description |
---|---|
namespace | The namespace used for functions in the datapack |
place location | An absolute position in the world to place the utility command block. Specifiy as 3 space-separated numbers |
spawn location | A position in the world to spawn the global armorstand. Specifiy as 3 space-separated components. Relative positioning (~) is allowed |
description | Set the description of the datapack |
generate cleanup | Generate a function which will remove everything created by the datapack |
MCC pipeline
There are 3 main stages to the compiler pipeline:
- Compiling: Converts source code into Command IR
- Linking: Merges one or more Command IR files into one master file
- Packing: Converts Command IR into Minecraft commands and creates a datapack from a datapack definition file
The MCC command takes a list of files, each code file goes through some number of transformations to end up as a Command IR object.
The linker merges each Command IR object into a single Command IR object. Any conflicts when merging will abort the compiler.
A datapack definition file is required for the packer. The packer takes the single Command IR object and generates the final mcfunction files.
To get an idea of the hierarchy of components, here is a diagram:
+---------------+
| |
| C Compiler |
| |
+-----------+---------------+
| | |
| CBL | Assembler |
| | |
+-----------+---------------+
| |
| Command IR |
| |
+---------------------------+----+
| |
| Command Abstraction | +--------------+
+--------------------------------+-------| Datapack |
| Minecraft Commands | Definition |
+----------------------------------------+--------------+
| Datapack |
+-------------------------------------------------------+
Running MCC
You will need to generate the standalone parsers (from Lark) using the ./build-parsers.sh
script.
If on Windows, run the python
commands in that script from the root of this project.
The Lark python package needs to be installed, pip
can be used on the requirements.txt
file. It is recommended to use virtualenv
.
Example: virtualenv env --python=python3 && source env/bin/activate && pip install -r requirements.txt
Once the parsers have been built, you don't technically need the virtual environment anymore. It can be deleted with
deactivate && rm -rf env/
Alternatively, you don't need to create the standalone parsers if you keep Lark available in the python environment.
MCC is implemented in python. Currently it is not bundled into an executable so it must be invoked using the python interpreter.
MCC is invoked by python mcc.py
(If python 3 is not your default python command then run python3 mcc.py
).
Command help text:
usage: mcc.py [-h] [-o outfile] [-E] [-c] [-S] [-dump-asm] [-O {0,1}]
[-dump-ir] [-shared] [--dummy-datapack] [--stats]
[--dump-commands] [--c-page-size SIZE]
infile [infile ...]
Command line tool for the Minecraft Compiler Collection
positional arguments:
infile Input files
optional arguments:
-h, --help show this help message and exit
Output generation control:
Options that control what stage in the output pipeline the compiler should
stop at.
-o outfile Set the filename to write output to.
-E Stop after running the preprocessor. Input files which
do not pass through a preprocessor are ignored.
-c Compile source files, but do not link. An object file is
created for each source file
-S Do not compile Command IR code. A Command IR file is
created for each non Command IR source file
C Compiler options:
Options specific to the C compiler.
-dump-asm Print The generated ASM code to stdout. Compilation is
stopped after this point.
Optimization control:
-O {0,1} Control optimization level
Linker arguments:
Arguments that control how the linker behaves.
-dump-ir Print Command IR textual representation to stdout after
linking objects, then exit
Packer arguments:
Arguments that control how the packer behaves.
-shared Include a symbol table and other metadata to enable
dynamic linking of datapacks
--dummy-datapack Don't write an output datapack. This can be used for
debugging with --dump-commands
--stats Print statistics about the generated datapack
--dump-commands Dump the commands to stdout as they are written to the
datapack
--c-page-size SIZE Size of the memory page to generate if using the C
compiler
MCC will dispatch a different tool depending on what the file extension of each input file is:
.cbl
Will compile the CBL source code file.c
Will compile the C source code file.asm
Will compile the assembly file.ir
Will read the Command IR file.o
Will load the pre-compiled Command IR object file.dpd
Will read the datapack definition file
Depending on the command line arguments, MCC will perform different tasks on each file. Some files are ignored if they are not relevant for the desired action.
Examples
Compiling examples/fib.asm
into a datapack fib.zip
:
python mcc.py examples/fib.asm examples/fib.dpd
Compiling examples/hdd_driver.c
into Command IR examples/hdd_driver.ir
:
python mcc.py examples/hdd_driver.c -S
Compiling mylib.cbl
into an object file, statically linking myprogram.cbl
with the object and examples/fib.ir
,
then using mydatapack.dpd
to create mysuperdatapack.zip
:
python mcc.py mylib.cbl -c
python mcc.py mylib.o myprogram.cbl examples/fib.ir mydatapack.dpd -o mysuperdatapack.zip