A toolkit for the Haverford Educational RISC Architecture (HERA) assembly language, comprising an interpreter, a debugger, an assembler, and a disassembler.
You can install hera-py with pip:
$ pip3 install hera-py
After installation, use the hera
command to run a HERA program:
$ hera main.hera
Enter the interactive debugger with the debug
subcommand:
$ hera debug main.hera
Assemble a HERA program into machine code:
$ hera assemble main.hera
Disassemble machine code back into the human-readable HERA syntax:
$ hera disassemble main.hera.lcode
You can also preprocess a HERA program without running it, to see how pseudo-instructions and labels are resolved to HERA code:
$ hera preprocess main.hera
HERA-C is the current HERA interpreter used at Haverford. It is implemented as a shell-script wrapper around a set of C++ macros that expand HERA instructions into C++ code, which is then compiled by g++.
hera-py improves on HERA-C in the following areas:
- Includes a purpose-built HERA debugger
- Concise and accurate error messages
- Ease of use
- Cross-platform and easy to install
- Configurable with command-line options
- Does not create temporary files
- Command name has six fewer letters than
HERA-C-Run
hera-py also supports several features that HERA-C does not:
- Setting registers to the value of a label
- Detecting stack overflow
- Multi-precision multiplication
- Relative branching by a fixed integer value (e.g.,
BRR(10)
) - Branching by the value of a register (e.g.,
SET(R1, 20); BR(R1)
) - Detecting invalid relative branches
HERA-C has a few features that hera-py does not:
- C-style #define macros (and more generally the ability to embed arbitrary C++ code in HERA programs)
Small to medium-sized programs generally run faster with hera-py than with HERA-C, while very large programs are typically faster with HERA-C.
Hassem is the current HERA assembler used at Haverford. hera-py has better error messages than Hassem, allows the user greater control of output (e.g., with the --stdout
flag), and fixes some Hassem bugs.
Thank you to Christopher Villalta for valuable feedback on early iterations of this project.