I’ve recently discovered a tweet that showed a Commodore 64C complete with green monitor and 1541-II drive, heavily dust covered, but still working and being used in a Polish Auto Shop. The picture itself was fascinating enough, but I’ve also discovered that it was a screen grab from a video – which I’m including here as well (had to replace the music with something royalty free).
In this screencast I’ll show you how to write your own INPUT routine in Commodore BASIC. This comes in handy when you want to reject certain keys from being used when asking users for keyboard input. In my example I’m going to allow all alpha characters (A-Z), as well as SPACE, RETURN and the DELETE key.
In this episode I’ll show you how to create the iconic flashing borders on Commodore machines. Back in the day, when the system was loading, this was a nice way to indicate that the computer is busy doing something rather than being dead.
I’ll show you the principle both in BASIC and in Machine Language on the C128. The VIC-II chip is the same on the C64 though, so this will also work on the Commodore 64.
The same approach can be used on the Plus/4, however the addresses for the border and background colours are different (decimal 65305, or hex $FF19).
In this episode I’ll explain the concept of Flow Control in Commodore BASIC. It’s kind of a video update of a post I did a while ago.
In essence, it means that we can tell the programme to take a different route in the code depending on a condition that’s met. We’ll explore the IF/THEN and ON… GOTO/GOSUB statements (available on all versions of Commodore BASIC), as well as the expanded IF/THEN/ELSE version (available on the C128 and Plus/4 only).
In addition, I’ll also show you how to use the BEGIN and BEND clauses that were introduced with the C128.
Did you know you can run Commodore BASIC v2 on your Mac and Linux systems as a scripting language? It’s true – thanks to the marvellous efforts of Michael Steil and James Abbatiello. They’ve adapted the original BASIC v2 as featured on the VIC-20 and C64 with additional routines so that it works natively on modern machines. It’s ingenious!
For those who don’t quite know what to do with it, here are some instructions that’ll help you get CBM BASIC up and running on macOS.
GitHub provides a convenient way to either clone a repository on your local machine if you have GitHub for Desktop installed, or you can download a ZIP file and unZIP it somewhere on your system. Once done, you should see a directory structure that looks exactly like the repo on GitHub.
You need a few utilities installed your your Mac to be able to compile files. Downloading Xcode will ptovide you with an excellent (and free) IDE as well as the command line tools needed to do that (gcc, make and several other goodies). You might be able to bring those components in using the Homebrew package manager.
Using your Terminal app, navigate to your unZIPped folder. It includes a MAKEFILE so there’s no need to mess with any parameters. Enter “make” at the command prompt, and after a few moments you should have a file called “cbmbasic” without an extension in the same directory. That’s the binary for your Mac.
To make it executable, we’ll have to tweak the file permissions – otherwise our operating system won’t be able to run it. We’ll do it using the chmod command:
You can test if it works by calling the binary without any parameters like this:
If all goes well you should see something like this:
For easy access, copy the binary over to your /usr/bin directory. That’s where common command line utilities are installed on Mac and Linux systems. The added benefit is that the path to that folder is already in your local $PATH variable, and as such you can simply type “cbmbasic” from any location when you’re on the command line.
Here’s how to copy the binary over (this will ask for your administrator password):
Using Direct Mode
When you run the binary it will bring up something like the standard BASIC command prompt we’re used to seeing on the Commodore 64. There are however a few important differences between a C64 emulator and this implementation:
this is NOT an emulator
cursor keys DO NOT work
all commands typed in must be CAPITALISED
Other than that, you can start typing like on a real machine. Be careful with certain POKE commands though as those may call system routines that might not be implemented.
LOAD and SAVE commands have been tweaked to reference the command line structure. They work just fine, but will load and save files in Commodore file format (which is not clear text ASCII – more on that later). For example:
The above displays the directory of the current folder, much like it was an attached floppy drive. How cool is that? You can reference folders using both LOAD and SAVE just as if you were on the command line.
You can also type in programmes and run them – however the cursor keys won’t work, so there’s no screen editing options. Programmes SAVEd from direct mode cannot be loaded from the command line, only from direct mode (i.e. when launching the binary without parameters).
To quit direct mode, hit CTRL+C.
Running Programmes from the Command Line
The real power of this implementation can be seen when we run files from the command line. Those files must not be SAVEd from direct mode, but instead are simple clear text files. The extension doesn’t matter, but for good practice, let’s store them as .BAS files (much like shell scripts are stored with the.SH extension, or PHP files would be stored with the .PHP extension… you get the idea).
You can write standard BASIC programmes use your favourite text editor (like vi from the command line), or try one from the “test” directory that’s provided with the repository.
Imagine we had a file called “hello.bas” that we’ve created with vi, looking like this:
To run our file, we can simply define it behind the binary like this:
This will greet us with the “Hello World” message on the command line.
Running Script Files
Alternatively, we can specify the full path to cbmbasic at the beginning of a file (she-bang notation) and run it just like any other script file. Observe please:
If the file was called “hello”, we’d need to change the permissions again so that the file can be executed:
Now we can run it like this:
Sweet – but I can’t work out how to compile this on macOS…
Fear not – I’ve got a macOS binary that was compiled on 10.12 “Sierra”. You can find it in my fork of the project. Check out the “binaries” folder.
Does it work on Windows? Or on Linux?
I’ve tested compiling and running this puppy on CentOS 6 and 7 with roaring success. The above steps worked fine for it, so I’m assuming it’ll work on other Linux flavours just as well (the beauty of portable C code).
According to the author, it works fine on Windows too – however I have no idea how to compile source code on Windows, so you’ll have to figure that out for yourself. I hear good things about Visual Studio – if I work out how to do it, I’ll add it to the “binaries” folder on my GitHub Fork.
Can I write my own extensions to BASIC?
Apparently so – check our Michael’s site and repo for details:
Right now, if you run SYS 1 from direct mode first, you can use the SYSTEM command (followed by anything you’d like to execute on the command line in “double quotes”) as well as LOCATE (followed by an x and y coordinate to place the text cursor) and WAIT.
Have fun hacking BASIC and letting it run with the blistering speed of modern CPUs 🙂
The Commodore 128 has a built-in machine language monitor which makes it ideal for ML development. However, most (or pretty much all) documentation on this subject is geared towards the Commodore 64, making it slightly difficult to get a head start in writing ML code for the 128.
Before I forget how to do it, here are a few pointers – courtesy of Jim Butterfield’s book “Machine Language – Expanded Edition”.
Let’s begin by typing MONITOR in C128 mode. It’ll take us to the machine language monitor. We’ll start our programme at $0B00. To begin assembling our code, we’ll type A 0B00 (A for Assemble), followed by these lines:
The MONITOR will turn this text into the output you’ll see in the screenshot above (the lines starting with a . dot). Here’s what this code will do when called:
First we’ll load the X register with a value of zero. We’ll use this register as a counter. In the next line we’ll load the accumulator with whatever is stored in address $0C10 plus whatever is stored in the X register. So if X has a value of zero, then the contents of $0C10 will be loaded. If X was 1, then the value in $0C11 would be loaded, and so forth.
We’re using this as ASCII representation of our text (Hello World in a box in this case). With JSR $FFD2 we’ll call a Kernal routine that prints a single character onto the screen. Now we’re incrementing X by one and ask if it’s 45 yet (CPX #$2D). This would indicate that we’ve printed all the characters we need. If that’s not the case, we’ll return to line 2 and keep printing. Otherwise, we’ll stop the programme.
Storing ASCII characters
You’d think it was possible to simply type in text in the MONITOR. But of course that would be too easy. Instead we need to grab one of those massive tables and hack in each character’s ASCII code in hex. How convenient!
Type M 0C10 (or whichever location in memory you’d like to store your text string at) and overtype the numbers at the start of the line, each one representing a single byte of our ASCII text. At the end of each line you’ll see what those characters look like when converted.
In my case it’s a total of 45 characters, beginning with a return, followed by the top of the box, HELLO WORLD, and the bottom of the box.
Running from the MONITOR
To start the programme from the monitor, we’ll type G F0B00. We’ll end up with a SYNTAX ERROR and back on the BASIC screen though due to the RTS command at the end of the listing. If we replace it with a BRK command instead, we’ll end up back in the MONITOR.
The important thing to remember is the five digit addressing mode on the C128 (i.e. G for GO, followed by F0B00). Our programme starts at $0B00 in memory, but to make it run properly we’ll have to specify which BANK it should be called from. Anything other than BANK 0 or BANK 1 is fine, otherwise we won’t reach the print routine at $FFD2. In my example I’m choosing F, but E would work fine too (as we’ll see in a moment).
Running from BASIC
Type X to exit the monitor and go back into the land of BASIC. First we’ll need to choose a BANK. We’ll have 16 to choose from (0 to 15), so perhaps let’s try BANK 15. Now we’ll need to type the start of our programme in decimal:
or we can use the DEC command to convert hex to decimal on the fly:
Saving the programme
From the MONITOR, we can save the programme using the S command. It needs to be followed by a name (in double quotes), followed by the drive number, memory start and memory end plus one byte – all separated by a comma. It’s probably easier to show than to write:
We’re saving more bytes than strictly necessary here due to the large gap between our code and the beginning of the ASCII string. Our string could go up to $0C3F. The last byte in $0C40 is NOT saved to disk (or tape).
We can do the same from BASIC using the BSAVE command (for Binary SAVE). The syntax is BSAVE “FILE NAME”, P1234 TO P5678. Sadly the DEC command doesn’t work inline with this command, which would make it extremely useful. We’ll have to convert the values into decimal manually instead.
Loading the programme
To bring our masterpiece back into the computer from the MONITOR, the L command works a treat:
From BASIC we can load the programme the usual way, making sure we load it with ,DEVICE,1 at the end. This is to make sure it is loaded into the same memory it was saved from, rather than the start of BASIC:
In this episode I’m adding statistics support to my previous lottery generator on the Commodore 64.
I’ll add an array that is updated if my supplied numbers have been matched, and how many times over how many draws this has happened. I’ll also add an option to pause the programme and display the statistics before random draws can continue.
When this app is run continuously it will collect statistical data on how many lottery draws are necessary to match all supplied numbers.
PS: By the time the video had uploaded, my emulator had drawn over one million sets, and none of them have matched my numbers 🙁
In this episode I’m amending my previous lottery number generator to take six lucky numbers from the user to match against the randomly drawn numbers from the Commodore 64.
This will allow us to compare what the computer has drawn to the user’s input, as well as keep drawing numbers until the user input comes up. It’ll be an interesting experiment to see how many draws that will take…