Cli Capabilities
Alongside the language itself, I'd like to include some information about the transpiler I wrote for it, because, of course, there's no point in writing a book about a compilerless language.
The only (probably forever) basm compiler is
available on github
under the retrospectively confusing name basm.
This tool doesn't just include a compiler,
it also comes with a bf interpreter to run transpiled basm directly.
Currently, the cli has two subcommands: compile and run.
Both of them are similar in that they both take in a file path and have some flags in common.
compile
The compile subcommand is the simplest way to use the basm cli.
Similar to tool like gcc, it will create a file in the current working directory named like the one passed in with the extension replaced with .bf.
File name and output path can be specified with the -o flag.
If compilation fails, error information will be printed to the terminal.
Flags
| Shorthand | Longhand | Description |
|---|---|---|
-o | --out <OUT> | The path to put the compiled file |
-p | --show | Print the transpiled bf script |
-u | --unoptimized | Skips the use of the inbuilt brainfuck optimizer |
-h | --help | Print help |
run
This command compiles then run the specified basm source file.
You can use the bf interpreter directly if you use this command the -r flag,
which specifies that the file is a bf source file, not a basm source file.
Unlike compile, this does not create a file containing the compiled bf.
If the compilation fails, error information will be printed to the terminal.
Flags
| Shorthand | Longhand | Description |
|---|---|---|
-c | --cell-size <CELL_SIZE> | Sets the size of cells in bits (only 8, 16 and 32) [default: 8] |
-i | --signed | Sets the cells as signed containing signed numbers |
-a | --abort-overflow | Aborts the execution of the program when a cell over/under-flows |
-t | --tape-limit <TAPE_LIMIT> | Limits the lenght of the tape in cells, aborts execution if it is reached |
-n | --number-input | Treats the input as integer numbers rather than characters |
-m | --number-output | Treats the output as integer numbers rather than characters |
-s | --single-input | Removes the bufferring of the unused parts of inputs to provide to them to later bf inputs |
-r | --raw | Interprets the file as brainfuck, skips the compiling process |
-p | --show | Print the transpiled brainfuck |
-u | --unoptimized | Skips the use of the inbuilt brainfuck optimizer |
-d | --dump | Dump the tape and tape pointer position to terminal once the program ends (includes by erroring out) |
-h | --help | Print help |
Bf Optimizations
basm applies some basic optimizations to the bf resulting from the transpilation process by default.
It can merge operators to reduce redundant use (e.g: +++- would turn into ++)
and it can reorder operations so that less tape pointer moves are used.
Overall, the purpose of the built in optimizer is to reduce the number of operators in compiled scripts.
By default, basm doesn't remove redundant operations and is generally quite stupid.
For example, an instruction like INCR 12 0;, which does nothing,
would still move the tape pointer to the 12th cell. This gets optimized out.
All optimization are made to have no impact on the resulting program, at a few exceptions:
- Willingfully moving the tape pointer before 0 will be optimized out if there is a cancelling move afterwards (e.g:
<<<<<<<<<<<<<<<>>>>>>>>>>>>>>>) - Any pointer moves that don't lead to other operations (at the end of the program) are removed
If you suspect that optimizations are messing with your program,
you can disable them with the -u flag.