Minimal BASIC interpreter for running MININEC
Project description
This is a BASIC interpreter for an old dialect of the language used in early IBM PCs – and, in fact even earlier incarnations of BASIC that existed on a UNIVAC (according to [1] the code there – which can be run with Yabasi – ran on a UNIVAC).
It is written in Python. I wrote this over a weekend to be able to run old MININEC code (the linked version of MININEC contains some small fixes). It has almost no error checking (it relies on the code being correct not trying to aid you in writing a new program in BASIC, I think the world does not need new code in BASIC). If you’re looking for a working BASIC interpreter, look at the pcbasic implementation, it is also in Python but faithfully reproduces the memory limitations of the machines at the time. And it seems to use single-precision floating point numbers. This is why I wrote my own interpreter: I needed to compare computations in double precision and I could not fit some examples into the limited memory of pcbasic.
I’m probably not going to put much work into improving this code, it has achieved the purpose: Running (and debugging in Python) old MININEC code to allow me to compare the computations in BASIC to my re-implementation of MININEC in Python, pymininec.
But in fact the code can now run more than the publicly available version of MININEC and implements some of the basic binary file I/O mechanisms of the time.
Changes
Version 1.0: More fixes
This version has a lot of fixes that go beyond just running MININEC, it can now be used to read binary files (MININEC used a pre-processor script to create a binary representation of an antenna geometry) and it can run programs without a line number on every line.
We no longer require line numbers on every line
Handle multi-line IF/END IF
Dynamic DIM statements
Binary File I/O with GET/PUT, note that a file is opened for binary if a record length is specified, this probably should change at some time to just convert everything written to a file from python’s string representation where necessary and open all files in binary mode. We use the default in Basic of opening a file read/write, so don’t be surprised if a file opened for reading which doesn’t exist is created with zero length.
Conversion of int and float from/to strings – in python the strings are represented as bytes objects – these are the functions MKI$, MKS$ and fixes to CVI and CVS
Bug-fix when a GOSUB is not the last statement in a list of statements (e.g. in IF expr THEN list-of-statements), this did not execute the GOSUB previously
Version 0.3: Fix rules for printlist
The syntax of parameters to the PRINT statement can either use a semicolon (or a comma) to separate expressions or just put the expressions together without a separator. The latter had some quite ad-hoc rules and I’ve removed many of them, the result removes a reduce/reduce conflict…
Version 0.2: First working (released) version
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
