I’ve given my whole public-facing infrastructure a belated spring-clean. I have now largely moved off my collection of VPS systems provisioned using Ansible, to a stack much more in-line with what I help customers build at my job. I’m now hosted on a managed Kubernetes cluster provided by DigitalOcean, and everything from this blog to the infrastructure automation is expressed as code in git repositories and rolled out using Concourse.
The whole thing has all been torn down and recreated from scratch several times a day with no issues so I’m feeling pretty pleased with it. I also have:
SSL Ingress through Traefik
Monitoring through Prometheus & Grafana
ELK stack for logs
Publishing this post is now a simple matter of committing the updates to git and then waiting a few minutes for my Concourse pipelines to pick up the change, build new containers and deploy. All of which means I now have far less excuses not to write more!
A few months ago, reading the news about Linux dropping support for floppy disks set off a whole bunch of memories and emotions around this long obsolete and “dead” data storage format. I hadn’t thought about or used floppy disks for around 20 odd years now, at some point I must have copied my last disk but I honestly can’t remember what it was or when. The floppy disk was one of those rare things that became part of your everyday life like your keys or school pencil-case. It was something from childhood and my early career that was everywhere: Scattered through your school bags, coat pockets and occasionally – if you were feeling organised – a disk box. And now, apart from a “save icon” meme, it’s gone.
For various reasons (mostly because I’ve just become a Dad!) I’ve been on quite the reminiscence trip recently and I decided to get reacquainted with this old and quirky relic of my past. Here’s the first of my disk boxes now, filling up quite nicely with little 3.5” slices of nostalgia after a few day’s worth of converting my ADF collection back into their original physical form:
I’ll cover the technical details of the “How” as per my usual Amiga blog posts but I want to talk a little more about the “Why” aspect. Be warned, I’m going to ramble on slightly more than usual as I go on a meandering stroll down Nostalgia Avenue.
I won’t deny the convenience of tools like WHDLoad but having nearly every single Amiga game and demo available at the click of a mouse has resulted in “the paralysis of choice”. I liken it to having a Spotify subscription: I have access to pretty much the entire recorded output of every band, ever – but there’s just so much choice I usually end up sticking to one or two playlists or putting it on random shuffle, invariably getting bored halfway through a song and endlessly skipping tracks until something catches my attention.
There really is something about the act of having a sort through the disk collection and picking something out akin to the ritual of pulling a vinyl record from it’s sleeve and placing it on the turntable. And the noises! That “schonk.. chonk…” as the disks slid in, followed by the “chnk…chnk…chnk…ddddddrrrrr..chnk…chnk” noises… So satisfying!
All these years on and it seems stranger still that this literal icon of obsolete technology once seemed like the future to me. Coming from a cassette tape-loading Sinclair ZX Spectrum, when I was younger the floppy disk was symbolic of the then next-gen systems like the Amiga. I started collecting and trading games and demo disks with school friends in anticipation before I even got my first A600. After all, the floppy disk was literally the first thing you saw when you switched on your Amiga, displayed right there on the boot screen:
Then from Kickstart 2.0 onwards we got the funky animated version:
Demo groups parodied it and riffed on it:
Whilst thankfully a damn sight faster than loading from cassette tape, most games would have a screen like this somewhere to remind you that they were actually doing something and not just aimlessly clicking away for several seconds:
And of course, without the floppy, we would never have been “treated” to the cultural milestone that is MC Double-Def D.P. performing “Don’t Copy That Floppy”. This is possibly the most awkward 90s thing I have ever seen – and that’s coming from someone who once wore a shell-suit, Global Hypercolor t-shirt and Hi-Tec basketball boots at the same time (yeah, it was a different era…):
If you’ve never seen it before, I challenge you to try and make it the whole way through. It’s just… it’s amazing.
There were other rituals associated with the floppy too; when I recently purchased my first new packs of disks in well over two decades, I realised how much I had missed the ceremony of labelling them! Here’s my fresh supply, halfway through being stickered up:
I remember there was something just so satisfying about lining up the labels just right and of course, deciding when a particular disk was important enough to warrant the permanent pen instead of having a tentative pencil scribble on the label. Or even, for those select few disks you made for your closest friends, you’d decorate the label with some home-made artwork in the same painstaking way usually reserved for your most important mix-tape covers.
Even those favourite disks occasionally had to be recycled; We usually ended up running out of blank disks at some point and then a friend would come over with something so cool you just had to copy it and then…. Well, time to pick a sacrificial victim and write over it. I’m sure we all had disks that ended up looking like this:
Which led to a fascinating geological-type effect, where the strata of previous games and demos were left exposed but scribbled out from your history of interests. You could see what provenance that disk had, remember gaming sessions of old and look back in time like some sort of 16-bit archaeologist sifting through ruins. Magazine coverdisks usually suffered this fate with me. As a subscriber to various Amiga magazines I was guaranteed a monthly supply of new disks which I could re-label and put back into use by the old “hack” of taping over the write-protect hole.
One of the interesting side-effects of floppy disks being an actual, physical medium is that you needed some way of obtaining and exchanging them. This in turn I think really helped the early computing scene become so much more personal than the modern era where everything is a quick download away.
Aside from the tradition of visiting your friends house and having a good old rifle through their disk-box to find things of interest, there wasn’t an easy way to get hold of new software like there is now. Back in the day, no-one was “online” and very few of my friends had modems. Those that did usually also had PCs which couldn’t read or write Amiga floppies. We therefore usually couldn’t simply download the software we wanted, which led to the proliferation of PD Libraries which advertised their wares in most magazines of the day. A typical example is this full-page advert from “17-bit Software” (which was interestingly enough the roots of the legendary Team 17 games publisher which brought us the likes of Alien Breed and Project-X)
Flicking through the back-pages of Amiga Format and CU Amiga was a regular past-time for me; I’d carefully read all the adverts, circling the titles that sounded the most interesting and then go and pester my Mum to phone them up and pay using her card over the phone. I’d then wait in excitement over the next few days for the postman to deliver the disks – it was always a good day when just before school, one of these bad boys would land on the doormat:
Later on, when I eventually got involved in The Scene, floppies brought us together in a way I think is now sadly long-gone. Because we traded disks with each other – and groups typically had members whose sole “job” in the group was as a “trader” or “swapper” – we also wrote to each other. As a particularly awkward teenager (with a bunch of the usual teen issues and a healthy side-order of angst), some of the closest friendships I formed during those years were with the other members of my group. Alongside the floppy disks we’d write long, rambling letters to each other full of everything and nothing. We’d fill our envelopes to each other with “Jiffy Junk” – little trinkets we’d collect and swap: Kinder surprise toys, trading cards and the occasional mix-tape. As the splash screen from an old disk magazine said:
Even though I think I only ever met one of the guys from my group in person, I still think about them often to this day and wonder what they’re all doing now.
As the Amiga scene gradually died and the BBS (and later, the Internet and FTP sites) became the primary means of obtaining software I think a lot of those friendships were broken. Even though The Scene is still a very close-knit community, I’d be interested to know whether those same bonds exist today when most communication is probably electronic in nature.
Back to the floppy
So, now down to business (skip to the “But seriously, why?” section below for more of my musings if you want to gloss over the technical details). My A1200 had been fitted with a HXC Floppy emulator: a nifty little gadget that loads ADF files from a SD card and plugs into the floppy socket on the motherboard. I’d installed this years ago for the sake of convenience and is what I had been using for a “floppy drive”. This had to be disconnected and I also had to make a cardboard insulating cover to shield the Indivision AGA DVI card which sits right underneath where the new drive was going to sit.
I purchased a refurbished floppy from amigastore.eu which I then installed :
Fortunately it sat secure enough with all the other cables around it, but it should have had a little bracket holding it in. It’s working fine now though without it and as it was a very tight fit with all my other expansions, I don’t much fancy opening it all back up any time soon! After a little fiddling with power supplies and working which way round the ribbon cables were supposed to go, it started up and proceeded to chunter away with that regular “click” of the Amiga waiting to detect a disk being inserted. Hurrah!
Now, to feed it.
There’s two types of floppy in circulation you can use with your Amiga. The first type is the native Double Density (DD) type which on the PC could support 720Kb of data and on the Amiga 880Kb. Although later Amiga models did sometimes come with a High-Density drive, this DD type was by far the most common type of disk in circulation at the time. They’ll work without issue in any Amiga drive but they are also the hardest to come by these days; the only reliable source for new DD disks I have found so far is from sellers on eBay. There are also plenty of “bucket of random magazine coverdisks and pirated games” auctions as well if you want to get really cheap.
The second type is the far-more common High Density (HD) type that most PCs used with a capacity of 1.44Mb. Again though, as they are no longer being produced they are now relatively expensive. You can still buy these brand new though, on sites like Amazon but there is a draw-back. They use a completely different magnetic medium than the DD type with different characteristics and as such they may or may not work reliably.
I tried 3 different batches bought from Amazon with mixed results. I bought a pack each of rainbow-coloured and plain black disks manufactured by Verbatim, Plain black disks from Sony, and a box of mixed-colour disks from 3M. In order to get these working I discovered the most reliable method was to use XCopy to format them several times in my external Cumana CAX354 drive.
I bought this drive from eBay as I wanted to recreate my then-pimpin’ dual drive setup I had on my first A1200. I think the write mechanism in this drive is somehow “stronger” than in my internal drive as I had a lot less issues once I used this to do the initial formatting. Here it is crunching through a X-Copy (Yeah!) format and verify:
Once I’d formatted and verified the disks each time, they usually worked OK but on average there seems to be 1 or 2 “bad” disks in each batch that either refuse to be formatted or eventually start throwing CRC errors.
I also needed some extra software tools to convert things back and forth. As my A1200 is connected to the Internet I went over to Aminet. There are many similar tools there but one I kept seeing recommended was TrackSaver GUI which could write all my ADF files back to a real floppy. It has a very simple GUI which takes care of reading and writing floppy images:
As well as ADF files it can also handle DiskMasher DMS archives which were commonly used on the Amiga back in the 90s. I remember this output was a common sight on magazine coverdisks and BBS downloads:
TURBO GENERIC, indeed. One last thing to consider would be a virus-scanner of some type. The Amiga bootsector virus was quite the thing back in the day, and it still persists on some ADF sources or old floppy images. It’s rare but it’d be a real shame to screw up a system by doing the digital equivalent of re-introducing smallpox. Again, Aminet has plenty to choose from.
But seriously, Why ?
I’ve been thinking a lot recently about what it is exactly about the Amiga that captures me in a way that no other computer ever did. Part of it is I think because the Amiga was the last system that was simple enough that one person could reasonably hold the entire machine in their head.
It was complex, true, but with enough studying and time you could know pretty much everything about every single bit of the Amiga’s memory space, workbench, the custom chips and so on. That’s what makes it so special and endearing.
That simply doesn’t exist on other platforms. I mean modern systems are amazing – and I’ve worked on some serious big-iron hardware in my time – but you could spend an entire lifetime now simply becoming an expert in one tiny piece of the whole system. Modern computers just don’t have the accessibility that the Amiga had, and it straddled the divide perfectly between being simple enough to know it and love it, yet powerful enough to do things you’d never dreamed possible.
But I would also suggest that an even more important aspect is the community that grew up around it – which was made possible in a big way by the humble floppy disk. It enabled a whole generation of bedroom coders to reach out to the world, sending their software around the globe and making real connections. Like the bootleg tape scene of the 80s we traded disks, wrote letters, made friends and connected in a way that now, like the floppy itself, seems antiquated. But it was a lot of fun while it lasted!
When the Amiga was introduced in the mid 80s, pretty much all displays available to consumers were 4:3 ratio TVs, and you plugged your computer into your TV via a RF modulator box. Later, dedicated monitors like the Phillips 8833 Mk2 (which I originally had as a teenager in the 90s) became available and these offered a much improved, sharper image but were still in the 4:3 ratio as this is what the native Amiga, PC and gaming console screenmodes had been designed around.
Nowadays however (with the odd exception) all commonly-available displays from TVs to flat-screen monitors are all in wide-screen 16:9, 16:10 or wider ratios. These displays tend to make everything designed for a 4:3 ratio look stretched and a little blurry. This is because the native screenmodes of the Amiga – such as 320 x 256 for most games and demos – cannot be expressed as even, equal fractions of modern widescreen resolutions. So you either end up “cropping” the display leaving black borders down the side, or you attempt to scale it up but end up with rounding errors and “pixels” that are wider than they are tall.
For reference, here’s a cropped photo I took when I first got my A1200 working again:
In this picture, I am using a standard modern widescreen monitor with “FHD” resolution of 1920x1080 pixels. It’s connected to the Amiga using an Indivision AGA board, and while I sort of adjusted to how it looked, it never felt right to me because of the squashed and stretched feel.
I can’t say exactly what triggered all this off, but I realised a little while ago how unnatural this looked and how much better screenshots always looked when presented in the correct ratio. I then remembered years ago that before everything went wide-screen, “square” 4:3 ratio monitors were pretty much all you could buy, especially the early wave of flat-screen displays. So I went searching through eBay and various other second-hand sites but I finally found a site selling brand new, unused 4:3 ratio monitors – it seems like there’s still a demand for the old ratios which is great news for us Amiga and other retro-computing enthusiasts.
I picked out the cheapest option which was an ASUS VB199T-P, a 19-inch SXGA LED IPS monitor with a native resolution of 1280x1024 pixels. This works out to be the perfect resolution for me as all the Amiga screenmodes I tend to run can be expressed as even fractions of this.
For example, my Workbench is 640x512 pixels in 256 colours which for my setup (50Mhz 030 with 16Mb RAM) is a good balance between speed and display quality. This means that every pixel in this screenmode can be displayed by the monitor as a 2x2 pixel square (640 x 2 = 1280, 512 x 2 = 1024). This can be seen most clearly when you look at an extreme closeup of the new display:
It’s perfect! No stretched, blurred or odd-shaped pixels, just lovely crisp and clear squares. This is what my Workbench environment looks like now:
Compare this with the photo at the start of the post; the effect is most pronounced in the text underneath icons, and in the EaglePlayer (music player that looks like a HiFi) window. Total “night-and-day” difference and I was blown away by how much better everything looked, not to mention how it felt so much more like my Amiga of old.
This monitor also works out wonderfully for most games and demos as these are usually (in PAL mode) in a 4:3 ratio of 320 x 256 resolution which works out as “pixels” of 4 x 4 on the Monitor. Demos in particular looked fantastic and it was great to see them as intended once more:
I got carried away and ended up working my way through all the classics, before eventually re-watching some true vintage gems from the late 80s and early 90s:
At which point it was getting dark outside and I realised I should probably call it a day and go to bed! In short, this is easily some of the best money I have ever spent on the classic Amiga and has totally changed the feel of the system for the better. All those classic games and demos now feel brand new to me once again; If you’re currently using a widescreen monitor on a classic ‘miggy, I cannot recommend enough trying out a 4:3 ratio screen. If you don’t want to go for a new one, you can probably pick up one for nearly nothing at your local second-hand store or website as they’re mostly seen as obsolete these days. Far better they end up in the hands of retro-computing fans than landfill!
Until next time… I have to go and have a couple of rounds with Project X now :)
It’s been a busy few months here, but I’ve still found time to enjoy my Amiga systems. I’ve been grabbing the odd hour here and there to continue my efforts setting up an Amiga development environment and “dip my toe in the water” again. My setcmd utility is progressing nicely and I’ve learned a lot about the tools like AmigaGuide and the Installer that I used on my A1200 back in the day. I thought since it’s been a while, I’d write a quick “brain dump” post and cover two of the big things I’ve finally got sorted out: Picking an editor, and getting Git working so I can publish and share my code.
There’s a huge selection of code editors available for the Amiga ranging from simple text editors to full-blown IDE systems, and everything in between. I started off using the built-in editor MultiEdit which was installed as part of the Enhancer Software pack from A-EON. This is a bare-bones editor but does have some nice features like the ability to open multiple files and so on – think of it like a “Notepad” on steroids:
I did miss syntax highlighting and other advanced editing facilities, though.
At the opposite end of the spectrum is CubicIDE. This is a full-blown IDE based around the GoldEd editor. Although it’s 68k-native software, it works perfectly under OS4 thanks to the built-in emulation in OS4. I purchased the full version which arrived on a CD-ROM and I also intend to install this on my A1200 when I get around to starting the port of my proof-of-concept setcmd shell script into Amiga C.
CubicIDE is very obviously geared mostly towards C development and includes a whole host of features around that but is still quite usable for shell programming. It also understands Installer scripts and provides some useful features, such as the ability to add commonly-used snippets such as dialog boxes and so on straight from a menu:
For my current stage of development though, it was overkill for editing and managing a shell-script based project. The syntax highlighting for AmigaDOS scripts was also a little basic, it only really seems to differentiate between comments, strings, and “everything else”. I’m still probably going to use it a lot later on when I get back into C development, but for now I needed something with more features than “MultiEdit” but not quite as advanced as an IDE.
OK, so some might say it’s “too Unixey”, but (Fun Fact Time!) VIM was actually first developed to bring vi to the Amiga. It’s been my editor of choice for years now due to my work with Unix-based systems and when I first got my X5000 I did try an old port of VIM. But I quickly discovered when run from the AmigaShell the terminal emulation wasn’t really up to scratch. I couldn’t get full-colour syntax highlighting working and various things just didn’t work as well as they do on other platforms, so I’d abandoned it as an option.
Back in May though, an updated version of VIM with a MUI user interface was uploaded to OS4Depot. This was an absolute game-changer for me – with a proper GUI and decent colour terminal emulation, VIM is now as usable (or not, depending on your point of view!) as it is on any Unix systems. It also feels like home to me and I can fly around it very efficiently.
Customising it is pretty much the same as it is on any other system. The documentation says that when you create a HOME: assign, VIM can look in there for a .vimrc file. So added the following in my S:User-Startup that points this to my Work: volume, which is essentially the equivalent of a Unix $HOME directory anyway:
ASSIGN HOME: Work:
With this in place, I could drop in my usual .vimrc (although I was sad to see that the netrw file browser is not supported under Amiga-like OSes) and I was off! It supports highlighting of pretty much any file I throw at it, although it doesn’t auto-detect AmigaDOS scripts. To resolve this, I enabled modelines in my Work:.vimrc file with the following additions:
And then added the following modeline near the top of my setcmd script:
; vim: set syntax=amiga ts=2 expandtab:
Here’s what a typical development session looks like for me now:
I have Vim running in graphical mode with a couple of tabs open, an AmigaShell session for testing and debugging and of course, AmigaAmp playing some classic MODs in the background. Bonus points to anyone who recognises where the currently playing track comes from ;)
I also spent some time making a nice dark colour theme for my AmigaShell to match the Gvim theme; while I did love the default grey and blue look of the shell for sheer retro appeal, I do much prefer working with this setup.
By the way, you may notice the background is slightly different in each of these shots – the little light-blue “bubbles” are actually drawn by CANDI which is part of the A-EON Enhancer Software pack. I remember the first time I clicked around some of the presets, I didn’t like any of the backgrounds so never really played much with it. But I then discovered that you can add your own backdrops, so now I have my favourite Amiga Keys wallpaper with lazily-drifting spheres over it. It’s really quite hypnotic, and very cool to watch!
So that’s what I’ve ended up on for now anyway. I’ll probably change my mind a dozen times in the next few months but for now I’m very happy with my setup and workflow :)
Last time around, I had experimented with the sgit client to try and make use of Git. Sgit is available from OS4Depot (link below), but there’s now currently an effort by another developer to update it and then merge the changes back into the upstream project. If you’re interested, I suggest you subscribe to this thread on Amigans.net.
My goal with using sgit was really two-fold: I wanted version control for my own use (it makes development so much easier when you can track your code changes) and I’m already familiar with Git due to my work. But I also wanted to publish the source of my projects somewhere like GitHub so that they’re available to others. I feel very strongly that as much software as possible for the Amiga should be open-source by default, as there’s so much great software that has been published without the source available and then later abandoned. Which means that it’s effectively frozen in time and can never be updated, so we’re either forced to use old software or constantly re-invent things when a working solution has already been found. In an eco-system as small and fragmented as the Amiga Next-Gen platforms, this is ever more important. Anyway, I’ll get off my soapbox…
The good news is I got everything to work but it took a little trial and error on my part. I finally got it all playing nicely together, so for the benefit of others, here’s my steps:
Firstly, there’s a couple of things that need to be setup in the Amiga environment. I downloaded sgit from OS4Depot.net here and unpacked it into my Work:c directory. I do this to keep everything I download and add to the system separate from my System partition, so in my S:User-Startup I have a line which adds this directory to my path:
PATH Work:C add
As sgit uses a statically-linked (and somewhat old) OpenSSL build, you’ll need to set an environment variable telling it not to verify SSL certificates. I did this by adding the following line to my S:Shell-Startup script as I only ever use sgit from the command-line:
setenv GIT_SSL_NO_VERIFY 1
Not great from a security standpoint, but I’m just happy to have a working Git client at this point!
With those things in place, I was good to go. I did however discover that I couldn’t seem to create a new, empty project with sgit init and push it to a GitHub remote; each time I tried, CPU usage shot up to 100% and my system locked up. No idea what was going on there, but a workaround is to simply create a new repository through GitHub, then you can pull that and add to it once it’s been cloned to your system. For example, after I created a new repository for SetCmd, I cloned it:
I then get prompted each time to enter my GitHub username and password – which turns out is not actually a password, but a developer access token. You’ll need to create one of these and then keep it safe somewhere as you’ll need it each time you push.
It would be nice if I could somehow cache this, but looking at the source for sgit, it looks like it’s currently not supported. No big deal though as it only happens when I interact with the remote end, and it looks like once I get a cross-compiling environment setup it would be fairly easy to add support for e.g. reading a token from an environment variable. I’ll see if I can do that at some point to help out with the project.
Here’s an example run from the AmigaShell showing some of the other sgit commands such as status and diff:
And the final proof – this is what I see in my browser from my Mac!
This is awesome. Once I’ve added a decent README.md and tidied up my code a bit, I’ll make the repository public and look at making a first proper release. Exciting times ahead!
To wrap up for now, I just want to say that modern developer tools such as a Git client, IDEs and editors are a real necessity for any platform to survive so I am very grateful to everyone working on these projects. Big fist-bumps to all involved in making all this possible :)
My A1200 has been on the “repair shelf” since September last year, when it suddenly stopped working. It was quite hard to diagnose the exact issue as the combination of my Indivision AGA and monitor meant that it never synced up in time to see any error message, and I only got the very briefest flash of colour before the display went black. I thought I saw a purple flash a couple of times; this colour doesn’t appear in any of the diagnostic charts I could find, but it may have been my monitor’s attempt at displaying a red flash. In which case, it could be the Kickstart ROMS at fault, but I couldn’t be sure. I had tried replacing the ROMs with another old set of 3.1 chips, but there was still no improvement.
I tried disconnecting all expansion cards, accelerator and so on, but it seemed to be “bricked”. I pretty much stopped working on it when I got my X5000 but always wanted to get the A1200 back up and running; Emulation is great these days, but there’s no substitute for the feel and “soul” of an actual classic system.
Anyway, I recently ordered a set of 3.1.4 ROMs, mainly because I wanted to add OS 3.1.4 to my set of emulated systems on the X5000 but as I had ordered the physical chips, I decided to try replacing the ROMs on the A1200 with this set as well.
Installing the chips was quite straightforward but I strongly recommend the use of a chip replacement tool to help you. As I was pretty sure the main board was somehow dead, I didn’t use too much care and I just used a flat screwdriver to lever the old chips out. This bent and damaged the pins so they would have been unusable if I decided to swap them back in. Here’s what the new ROMs looked like inserted into the mainboard:
You can also make out the CF card holding my 3.9 OS and software, an Indivision AGA underneath (the red board) and a Blizzard 1230 Mk IV clocked at 50Mhz with 16Mb RAM. I put everything back together and not expecting much, flicked the power switch.
And, oh my word…. IT WORKED!
All I can guess at this point is that I somehow ended up with a couple of pairs of faulty ROMs before I tried the 3.1.4 set. It’s fantastic to have the old beast up and running again, and I spent an afternoon rediscovering all the projects I had been working on, got it back online with a PCMCIA network card, listened to some classic MODs and played some classic games.
The X5000 was not happy
The next morning though, disaster struck. I went to boot up the X5000 so I could get them networked together and share files, and….
Yeah, that doesn’t look good. Every time I pressed the power button on the X5000, the case fans briefly twitched but that was it, no further activity and the system wouldn’t power on. It was almost like it had died in protest at the A1200 being resurrected – like “Highlander”, There Can Be Only One!
I contacted AmigaKit where I purchased the X5000 from as it was still under warranty and they helped me through a very thorough set of trouble-shooting steps. Big thanks to Chris at AmigaKit who responded to my emails, and as it turned out was the person who originally put the X5000 together for me when I ordered it.
I guess this is one of the advantages of the Amiga scene being so “niche” – there can’t be many places where you can talk to the person who built your system and have them help troubleshoot it with you! Here it is with the sides removed while I was working on it (the inflatable Boing Ball came with my 3.1.4 ROMs):
We ran through what the status LEDs on the motherboard were doing, checked the SD card holding the BIOS, removed all expansion cards and also tried disconnecting the front-panel I/O and starting the system by connecting two jumper pins (to rule out a simple button failure or short in the case). The final step was to try a new PSU, and it finally span back into life!
A big relief, and I want to again thank Chris and AmigaKit for their help, and also to everyone who replied to my thread at AmigaWorld.
Together at last
And to round off this exciting few days, here’s what I had been hoping for the last few months – the “next-gen” and “classic” worlds side by side:
As I mentioned in my last post, I’ve been developing a new “SetCmd” tool for my Amiga systems, both for my own practical usage and also to serve as a re-introduction to the Amiga developer environment. This series of blog posts will cover the progress of this tool, as well as explore the challenges and technologies from my perspective of a returning AmigaOS fan.
I’ve learned a lot over the past few days and even though the tool is currently very much a “prototype” AmigaDOS script, it is working well enough at this stage that it has really helped my day-to-day activities at the Amiga shell. There’s a few screenshots at the end of this post showing my current progress, but to start off with I wanted to share my first “brain dump” covering some of the things I have picked up and/or re-learned :
As I want to make this tool as Amiga-native as possible, I also want to provide documentation in AmigaGuide format (an interactive hypertext-based system, somewhat similar to HTML), and an installer using the standard Amiga Installer utility. I decided to tackle the installer script first, and have already learned quite a bit.
There is apparently a new Python-based installer system present in AmigaOS 4 which can create much nicer and more modern-looking installer wizards (such as the SDK installer – I looked at this amongst other things in another post), but I chose to stick with the “classic” installer tool for several reasons:
At some point, I would also like to provide this tool for classic AmigaOS 3.x so I don’t want to make too many choices at this point that limit my options.
There are many Installer scripts to use as examples.
Pretty much every Amiga user is familiar with the Installer tool.
I remember it well from my classic Amiga days and always wanted to learn it!
I started off by examining some existing Installer scripts to get an idea of what the language looks like. It’s an odd sort of LISP-based system, which uses some patterns I found extremely jarring at first. For example, in most languages I am comfortable with, comparision is done like if thing == value. In the Installer language, it looks instead like (if (= thing value)). Still, once I’d got used to that it wasn’t too bad, and I discovered the language is pretty full-featured with procedures, control statements and lots of useful helper commands/functions.
As an example, here’s a block of code which copies the setcmd program file over to a previously created directory, the value of which has been assigned to the #dname variable:
And for running commands, you can use the run command, along with cat (short for “concatenate”) to build up the command string :
(run (cat "C:MakeLink FROM " #dname "/cmds/setcmd/release TO SETCMD:setcmd SOFT"))
I initially found it very difficult to find any definitive guide for this language as nothing was showing up on web searches. After struggling through with trial and error and only example scripts to help, I remembered that this is Amiga-land so there’s probably documentation in a native format. And sure enough, searching Aminet brought me to the Installer dev package. It dates from around 1996 which is shortly before I first left the Amiga scene but still seems to be a relevant source of information. The supplied Installer.guide is a great resource, and you can see it open below along with my Installer work in progress:
I also discovered that you can pass useful information and configuration which controls the Installer tool through standard Amiga tooltypes:
The more I get back into the Amiga, the more I appreciate these details and features that never seemed to make it into other systems!
As I use version control (usually git) for all my other projects on Unix, Mac and other systems like Haiku, I wanted to do the same for this project. However, this is something I’m kind of stuck with at the moment, so I’m curious what other Amiga developers use. Simply put, the only tools I could find (after an admittedly short search) were either very old or incomplete. There is for example, a port of Subversion to AmigaOS 4, but I consider that a technical step back and the port is already 10 years old.
My options for using git seem quite limited – all I could find was sgit. This is sort of working OK at the moment, I have managed to initialise a local repository and commit code but various things don’t seem to work quite how I would expect. I still haven’t worked out how to add and push to an origin at Github.com, for example. Diffs and .gitignore files also don’t seem to work although that’s probably something I’ve done wrong, and many of the commands don’t accept the same arguments as I’m used to.
Still, I’m extremely grateful for it as having a limited SCM system is infinitely better than the old approach of millions of files all called things like final_1.bak-v2-bug-fix_DO_NOT_DELETE which you can never remember what they are or why they were labelled like that!
I’ll keep investigating – one option might be to see if the bundled Python installation is up to running Mercurial. I once submitted a patch for this to work on Irix, so maybe in the future they’ll get another patch from The Weirdo Who Likes Old Systems ;)
This is more a “nice-to-have” but I admit to being stumped right now. The only way I have of developing this tool is by sitting at my Amiga X5000 which is fun (especially with AmigaAmp blasting some classic 90’s MODs in the background!), but it would be nice to be able to continue on with my work while sitting with a laptop in my living room. So far, the only route seems to be emulation on a Mac/Windows/Linux laptop which is not ideal.
I can’t seem to find a SSH or Telnet server for AmigaOS 4 (plenty of clients are available) which would have let me work from the shell remotely. There is a port of VNC, but using that for development work over Wifi is painful. I’d appreciate it if anyone has any insights – is there anything obvious I’m overlooking ?
Brief shout out to whoever maintains the AmigaOS Documentation Wiki – this has been an invaluable resource. I managed to get myself back up to speed with AmigaDOS after a multi-decade absence using the great documentation, particularly these fantastic guides :
AmigaDOS Commands – all commands described along with plenty of helpful usage examples and tips. The documentation for the list command is a revelation – I had always been mystified by that command back in the day!
Using Scripts – lots of great examples and tips on how to write scripts, accept arguments and so on.
AmigaDOS Advanced Features – more useful tips including a whole section on escape codes for colourised output. I remember working these all out by hand on my A1200 and working out what codes I needed to build up ASCII-art screens for my old scene group!
I saved all these guides locally as PDFs using the “Save as PDF” option from the Odyssey browser. I would have loved to have a resource like this back when I was starting out on my old Amiga, so again a big “Thank You!” to whoever is responsible for maintaining this wiki.
I am now able to use the setcmd running through itself to develop it. This has been very useful and has further validated my need for such a tool – take a look at the example below, where the setcmd installation is pointing to my development version in my Work: volume :
But now I want to switch to my release version, where I’ve installed it to my usual Software:Programs/ directory :
Note that the version number changes in the header of the command, and setcmd show setcmd shows the versions changing. Also note the liberal use of AmigaDOS Assigns – one of my favourite features of this OS, and one that I really miss on other systems. VMS had something similar with System Logicals, and of course Unix systems have symlinks but they don’t work in the same way, and aren’t nearly as useful or efficient.
My next steps are to finish the Installer script, then learn enough AmigaGuide markup so I can together a set of documentation for my tool. A few more bug fixes and general polishing, and I should be ready to release it to the world! This is just the start, however. I have more plans like supporting AmigaOS3.x – probably limited to OS3.9 as that’s all I have readily available at the moment but I do have a set of 3.1.4 ROMs being shipped.
I’d also like to try out different editors, and use my setcmd tool as a jumping-off point into re-learning C on the Amiga, eventually re-implementing it as a pure C program instead of the “prototype” script.
Part of the reason I started off on this journey was to develop my skills to a point where I can eventually start contributing more to the Amiga scene anyway, so hopefully learning enough C (and working out the build process, accessing Amiga libraries and so on) will get me to a point where I can help bring some more tools across to our platform.
Until next time…
=Haiku is a recreation of the BeOS, which I used extensively after I left the Amiga scene. I’ll write more about this OS in some other posts, but it’s nothing short of amazing what this small group of dedicated developers has accomplished. I’ve also been continuing in my mission to submit patches to open-source projects for obscure platforms ;)
I’ve been having a lot of fun with my X5000 over the past few weeks (more blog posts to come!) but I’ve been working on something recently that I wanted to share. I’ve been enjoying re-learning AmigaDOS and as an exercise for myself, set about building a tool I plan on releasing in the near future. Inspired by some Linux distributions’ “alternatives” system, It’s called setcmd (short for “Set Command”) and lets you easily and quickly switch between different versions of a command while in the AmigaDOS shell.
It’s been particularly useful for me as I experiment with different versions of the UAE emulator. I wanted to share a quick “sneak peak” at my work so far, but bear in mind this is all a “work in progress” and I have a lot more to learn! Anyway, here’s the “usage” screen that provides an overview of the program options:
And here’s a quick walkthrough demonstrating how I use it: First, let’s add a new command under setcmd control. I picked “UAE” for this test, as I frequently want to switch between versions while I’m testing compatibility with my various emulated systems:
I added the command, and now setcmd list shows that there is a new uae command available (note that the setcmd script is also managed by setcmd, so I can easily test new versions!)
A setcmd show uae shows that it’s just a “stub” entry for now; As no versions have been added, if we try to run it we just get a message telling us to add some versions.
OK, so let’s do that – I have two main builds of UAE I switch between: the system-provided one, and the new UAE-1.0.0 I downloaded from OS4Depot:
I’ve now setup two “versions” of UAE, one called system, and another called 1.0 – versions in setcmd can be any format so it’s easier for you to label them according to your needs. So, let’s see it in action! First, let’s try setting the version to “system”, and run the which command to verify that it’s pointing to the correct location:
And now, let’s switch it to the “1.0” version and again verify the path has changed:
There you have it! I can now easily and quickly switch between multiple versions of tools. It’s been a really fun experience diving back into Amiga development again, and I hope to release setcmd on OS4Depot.net soon. First though, I have to tidy it up, remove some bugs and I also want to learn the AmigaGuide and Installer languages so I can provide documentation and an installer to turn this into a “proper” tool. I look forward to my first proper Amiga software release in decades!
Happy New Year everyone! I’ve got big plans for my Amiga projects in 2019, but thought I’d start off the New Year with a blog post on a not-particularly “exciting” topic, but an important one nonetheless: Backups. As I am experimenting more with my X5000 and Amiga OS 4.1, I’ve been getting particularly “twitchy” that I didn’t have a solid backup/restore plan in place, particularly as some of my experiments will invariably go wrong and I’ll need a way to roll back my changes to a known-good state. I spent a few days researching and implementing a backup strategy that’s ideal for my needs and hopefully there will be something of use to other Amiga owners too.
My requirements were:
Create a spare bootable system partition that always has a working copy of Amiga OS 4.1 complete with all my settings and tools
Backup my system and other partitions to directories on another disk
Create archive files of these backups so they can be stored on another system
I’ll run through my steps for each of these areas in turn. Firstly, I wanted to create a spare bootable partition so that even if I totally screw up something in the AmigaOS boot process, I can still get back to a working system. To do this, I decided to use some of the spare space I had left unpartitioned when I set up my X5000.
Preparing the backup boot partition
The first step was to run Media Toolbox and select my internal boot device which is connected to the on-board p5020 SATA ports:
Next, I selected my boot device which is the 256GB Sandisk SSD installed by AmigaKit:
And here’s what my partition layout looked like, with a decent amount of space free for a secondary boot volume:
I know AmigaOS is a pretty light operating system, but I first checked how much space was actually in use on my current System volume using the AmigaDOS info command:
Not bad, just around 900Mb in use. I decided to create a 5Gb partition for my new boot volume; this was way more than I needed but then disk space is cheap, and it’s always better to have spare capacity in the future:
I named this partition DH3 as I’d already used DH0 to DH2 for my System, Software, and Work volumes. I then set the filesystem to SFS2 as that’s what I’m using on all my other volumes and it’s proven to be reliable so far. It doesn’t however have any recovery tools unlike NGFS, which means that a backup plan is even more important as I’ll have no choice other than to reformat and restore data if any corruption occurs!
The next step was to make sure that I had set the various options needed to allow booting from this new partition. I ticked the “Bootable” checkbox, and gave it a Boot Priority of -20, as that is lower than my current System: volume, which has a boot priority of -15. So this means it should only be picked automatically as a boot volume if my main system volume is unavailable:
With that done, I clicked on “OK – accept changes”, and back at the unit selection screen I clicked on “Save to disk” to write the new partition table to the SSD:
I then formatted the volume (using “quick format” – never do a full format of a SSD!) in Workbench, named it SystemBackup and I was ready to start copying data over.
Making the backup
I wanted to make the backups using something a little more efficient than the plain old AmigaDOS COPY ALL. On Unix systems, there’s a fantastic tool called rsync that can be used to archive directories by only copying over the changes between them. This means that the first time may take a little while as it has to do a full copy, but after that any subsequent runs will be much faster as only the delta will be transferred.
I tried a few of the available tools that can do incremental copies (as well as correctly preserve all the AmigaDOS file attributes) and eventually settled on the fantastic BackUp from OnyxSoft. It’s a 68k binary but works very well on both classic and NG systems – another great example of how AmigaOS 4 can run old system-legal classic software. It has a nice GUI for one-off operations and can also be scripted from the CLI (more on that later). I unpacked this tool into my usual Software:Programs directory, and ran the first backup manually. I did this by picking the Source and Target volumes and clicked the “Backup” button:
This started the backup running and produced some short output showing which files were being worked on:
And a short while later, the backup had completed:
I then ran the backup again to verify that incremental backups were working; sure enough, it completed in a fraction of the time and only transferred modified files.
Now I had my backup boot partition ready, I somehow had to find a way to boot into it. On my classic Amigas this would have been straightforward, using the “early boot menu” you can access by holding down both mouse buttons when powering on the system. There didn’t appear to be anything like this in the Uboot firmware on the X5000, though. The only options I saw on that menu are for booting into AmigaOS, MorphOS or Linux. The “Boot options” submenu just lets you pick whether to boot from HD, USB or optical media. So I knew I’d have to mess around with Uboot configuration directly, using the “Command Line” option from the boot menu.
Update : Trevor Dickinson of A-EON has provided a comment below that corrects my assumption above! There is apparently an early boot menu, you just have to press the mouse buttons together at the right time. Many thanks Trevor, for the update! My steps below are still valid, but it’s good to know about this extra environment. I’ll explore this in more detail later on…
The first thing I wanted to do was run a printenv command from Uboot and note the values. This is very important as it’s quite possible to make your system unbootable by changing things here, so having a record of settings is very important in case you need to roll back your changes:
The first parameter caught my eye – amigaboot_quiet. I had no idea what this did, but I hoped it would show some more verbose output during boot. I thought I might then see something in the boot logs which might point me in the right direction to set up multi-partition boot. So I set this value to N (hopefully to disable “quiet booting”) using the setenv command, and then ran boota to start the AmigaOS boot process off:
By sheer luck, this turned out to be exactly the parameter I was looking for! Now, when the boot process starts, I get a menu allowing me to pick a boot volume:
I was really happy that my first guess seemed to take me in the right direction :) I then simply used the cursor keys to move the selector arrow down to select the new backup boot volume, and pressed Return:
This started booting and loading kernel modules:
And then a few moments later, I got the OS 4.1 splash screen and then my Workbench desktop appeared. I celebrated briefly, but I quickly noticed there was a slight problem as it didn’t appear to have worked as I had expected. The system did boot, but when I checked my assigns I saw that SYS: (the boot volume) was still pointing to my System volume on DH0, along with all my other assigns:
This meant that although it appeared that the Kickstart kernel and modules etc. had been loaded from my backup partition on DH3, the boot had continued from DH0, run the startup-sequence from there and so on. So I was getting closer, but there was still obviously a last piece of the puzzle to put together.
I then went back to reading more about Uboot and the AmigaOS 4 boot process. I eventually stumbled upon a thread on the Amigans.net forums where someone was trying to do exactly the same thing as I was!
It turns out there’s one more important variable to set: os4_bootdevice should be set to auto. With this in place, the boot sequence will continue from whichever volume the Kickstart and Kicklayout file were loaded from. Otherwise, it appears the boot process will continue from whichever partition has the highest boot priority – which would always be my original installation on DH0:. I tried setting the two Uboot variables again before trying to boot from the new partition again:
I again picked the backup partition from the boot menu and this time, I had success:
You can see above that my SYS: volume is now pointing to my backup volume, along with all the other assigns. With that work verified, I went into Uboot for the last time, set the two variables I needed and then saved the configuration to the flash device with the saveenv command to make my modifications permanent:
Now, every time I boot I get the boot selection menu and can pick which partition to boot into.
Remainder of backups
Now that I had my bootable System: backup partition, I wanted to also create clones of my other volumes to make sure I have backups for my Software: and Work: partitions. For this, I just wrote a very simple AmigaDOS script (saved into my Work:c directory for all my personal scripts and tools) that runs OnyxSoft BackUp in CLI mode. It first updates my bootable clone, then clones everything (including the System volume again) into a Backups:Clones/ directory on a separate disk. This means that even if the SanDisk SSD holding my volumes dies, I still have a backup on another disk:
SET backup_bindir Software:Programs/BackUp
SET system_source System:
SET system_clone SystemBackup:
SET software_source Software:
SET work_source Work:
; Bootable clone
echo "Cloning $system_source to $system_clone"
$backup_bindir/BackUp SOURCE $system_source TARGET $system_clone BACKUP AUTOQUIT
; Backup clones on separate Backups: volume
echo "Cloning $system_source to Backups:Clones/System"
$backup_bindir/BackUp SOURCE $system_source TARGET Backups:Clones/System BACKUP AUTOQUIT
echo "Cloning $software_source to Backups:Clones/Software"
$backup_bindir/BackUp SOURCE $software_source TARGET Backups:Clones/Software BACKUP AUTOQUIT
echo "Cloning $work_source to Backups:Clones/Work"
$backup_bindir/BackUp SOURCE $work_source TARGET Backups:Clones/Work BACKUP AUTOQUIT
; Display backups usage
This took a long time the first time I ran it, but thanks to BackUp only copying changed files, it now takes only a minute or so each time I run this script.
I was now feeling a lot safer with my volumes all backed up, but I had one more step to complete, namely taking these backups and turning them into single-file archives which I could then transfer off my X5000 to another system, such as my home NAS and S3 for off-site backups. After all, it’s no good having all your backups on one system if that system completely fails or is damaged in some way. My years of running Unix systems professionally has taught me the “3-2-1” backup rule: A good backup plan must have at least 3 copies of the data, 2 of which should be on different media, and 1 should be stored off-site.
So, the final step was working out how to take these backup directories and archive them into single files. My first thought was to use LHA which is pretty much the standard Amiga archival utility. But there was a problem: As it turns out, most “classic” Amiga software is not large-file aware, which means that LHA could only create 2GB archives (which would have been unthinkably huge back in the day!).
This was OK for my system volume which was under 1GB in size, but it would not suffice for my Software and Work volumes which were much larger. I went searching for an alternative solution which could handle creating large files, as well as preserving all the important Amiga filesystem attributes. I posted a question over on the helpful AmigaWorld forums and quickly got a suggestion to try AmiDVD which sounded like a good tool to try as it’s bundled with AmigaOS 4.1.
I’m pleased to say this worked out very well, and was simple to use. I just used the “Create Image” tab, and tested it first by using my Work: volume which was around 40GB in size:
With this done, I then mounted the resulting .iso file and browsed it to verify that my files were all present and still had the correct filesystem attributes set. In the screenshot below you can see the ISO file mounted as a virtual drive, and you can see from the shell and Workbench window that the ISO file is around 40GB in size. I also browsed to the backup of my Work:c directory where I work on my own scripts and verified that the AmigaDOS filesystem attributes were still set correctly:
I then just had to re-run through the process again to create backups of my other volumes. Here’s the results, showing again the large filesizes created with no problem:
I could then transfer these files off my X5000 and onto my Synology NAS and other systems. I would ideally have liked a tool that could be driven from the command-line so I could include in my backup script, but AmiDVD is very straightforward and it doesn’t take much effort to click a few times to create updated backup archives. Still, I’ll continue searching and if I find something I can use from the shell I will update this blog post to cover it. Maybe mkisofs would be a good starting point…
Anyway, now I have a solid backup plan in place (and learned a lot more about my X5000 and the world of Next-Gen Amiga systems!), I can press on with having fun experimenting with my new system and I look forward to blogging more this year. Hope it’s a great year for all you Amigans out there and remember: Only Amiga Makes It Possible!
A little while ago, an updated port of the FUSE ZX Spectrum emulator was uploaded to OS4Depot. My first computer was a ZX Spectrum 48k, although I eventually ended up with a +3 model before I upgraded to my first Amiga. I was looking forward to getting this emulator installed so I could re-visit some of the classic 8-bit games I used to play; my favorite game is still Manic Miner although I never manage to get past the “Eugene’s Lair” level on my first attempt!
I hit a few stumbling blocks along the way though, so this “simple” exercise turned into a trip down the rabbit-hole which ended up teaching me a lot about my new X5000 system and AmigaOS 4. What follows is pretty much a blow-by-blow account of what I discovered along the way. As I’m still pretty new to OS4, I’m sure there’s probably a much neater solution to all this though, so any comments and feedback are all welcome!
Anyway, here’s the first problem I hit: I downloaded the ZIP archive for FUSE, unpacked it with UnArc and tried to launch it by double-clicking on the Icon. Instead of seeing the FUSE window appear, I instead got this error message:
I remember having this kind of issue all the time on my classic Amigas – a program would require some library that I didn’t have installed. Back in the day, if I was lucky I’d get an error message telling me what to download, or a note in a README guide somewhere. When that failed, I’d usually break out the fantastic and indispensable SnoopDOS utility to find out what a program was looking for. In this case however, it was clearly telling me what file I needed, but I had no idea where to install it. On my classic Amiga systems, standard Amiga libraries (with the .library extension) simply got dropped in the LIBS: volume, which was usually in the Libs directory on your startup disk. But I had no idea what to do with ELF-format shared libraries.
Anyway, I searched on OS4Depot for “libpng”, and found this package. Looking at the contents of the archive, I could see this contained a libpng16.so.16.34.0 file. This looked to be exactly what I was after! I downloaded the libpng archive, and unpacked it to a new directory in my Downloads folder (I really dislike archives that unpack into the current directory; it makes cleaning up and keeping track of where files came from so much harder). This left me with a bunch of library files and an AutoInstall file. I looked at this, and it turned out to be a simple AmigaDOS script:
Fairly rudimentary, but it’s clear what it’s doing: It first deletes any existing libpng-related files, and then copies things to the required locations, and sets up some symbolic links in SDK:local/newlib/lib. If you’re familiar with how Unix systems handle shared libraries, this should look pretty similar. I also discovered from that script that AmigaDOS can make use of symbolic links – I’m not sure when support for that was added, but I don’t recall it being around on my classic Amigas.
Only problem was, I didn’t have any SDK: assign, or any directory structure resembling the archive on my system. After some searching online, it appears this is created by the official SDK for AmigaOS 4, which can be downloaded from the Hyperion website. I picked the most recent version, SDK_53.30.lha (Software Development Kit for AmigaOS 4.1 Final Edition) and downloaded it.
After I’d unpacked it, I ran through the supplied installation utility, only changing the location to my standard Software: assign. I selected a “Full Install”, as this bundles a lot of very useful-looking GNU tools and utilities, as well as include files and compilers:
Definitely a lot to go exploring through on another occasion! I noticed that it had also modified my S:User-Startup file to make sure the SDK: assign is created, and it also dropped in another startup file that is run on boot:
Looking at SDK:S/sdk-startup, it appears that this sets up assigns and other things that are needed for the bundled compilers and utilities. In any case, I now had my SDK: assign, so I could go back to my libpng download directory and run execute AutoInstall from an AmigaDOS shell. This completed, and I could then see that the libpng files were copied to the correct destination under my SDK root. So, I clicked on the FUSE icon again, and…. same error: “Failed to load shared object libpng16.so.16.34.0”. Damn.
At this point, I discovered there was a SOBJS: assign which appeared to point to the System:Sobjs directory which also seemed to be full of .so library files. I decided to try and copy the libpng library there:
And I then re-launched Fuse. Now, I saw a different error – this time, complaining about not being able to find libz.so.1. One step forward, two steps back…. This libz.so.1 file however was in the SDK:local/newlib/lib directory, but I decided against copying everything in there over as it seemed pretty hacky, and a sure-fire way to screw something up. Plus, it seemed that there must be a way of adding this SDK lib directory to the system-wide library search path.
Now, on Linux or Unix systems, I’d have looked for a way to either re-compile Fuse with the correct -L and -R flags, or found a way of adding the SDK to the LD_LIBRARY_PATH variable. In this case though, I was totally stuck and in a very unfamiliar place. I shelved my “get FUSE installed” project for the time being, and instead took a brief diversion into getting QT installed. Quite by chance, I noticed that it added something very interesting to the end of my S:User-Startup:
Whoah. It looked like it was appending a path onto an existing Assign – I had no idea that AmigaDOS supported this! I’ve used the ADD syntax to the path command many times when adding additional directories full of tools, but never thought to try it with an Assign. It looks like it was added into AmigaDOS release 2, so it’s always been available to me ever since I got my first A600. Another new thing I learned!
So, I then added the following to the end of my S:User-startup:
; Needed for FUSE etc to pick up on new libraries
assign Sobjs: SDK:local/newlib/lib ADD
I rebooted, clicked on the Fuse icon, and…..(cue drum-roll)
Success! Although, I have to admit that at this point, I lost around 2-3 hours downloading and playing tons of classic games from the World Of Spectrum archives :)
However, this little distraction did teach me some new things:
Assign can take an ADD parameter which appends directories onto a logical assign.
The SOBJS: assign appears to be where programs look for ELF-format shared libraries on startup.
The OS4 SDK contains some very useful CLI tools and is worth installing on any Amiga OS 4 system even if you don’t intend to develop software with it.
Even after 30-odd years, there’s still something new for me to learn about classic AmigaDOS commands!
Finally, I’d just like to note that shortly after I solved the FUSE problem, someone very helpfully posted a comment which highlights an alternative fix – simply copying the libpng library to the FUSE directory. Although I wish I had thought of trying that, this little diversion proved to be a very useful exercise and I’m glad I got to dig a little deeper “under the hood” of my new X5000.
While I’ve been having a lot of fun with the new software written specifically for AmigaOS 4, the bulk of my software is still “classic” titles that used to run on my A1200. One of the first things I did when I set up my X5000 was to transfer my old Amiga’s hard drive over so I could continue running this library of software. I also wanted to set up an emulation of my A1200 so I can quickly launch a classic Workbench 3.9 session and pick up all my old projects and bits of code I’d written over the years.
Fortunately, the X5000 and AmigaOS 4 offers a variety of ways of running all your old software. For this blog post, I’ll break it down into four areas:
Running software from disk images (such as ADF, DMS etc.)
Running games and demos from WHDLoad
Full system emulation with E-UAE
Running 68k system software
The first “mode” I wanted to explore was running classic Amiga games and demos from ADF disk images.
I discovered that thanks to the pre-installed RunInUAE tool, this was about as easy as it gets – you simply double-click a supported disk image filetype such as an ADF or DMS image, and RunInUAE then boots a pre-configured fullscreen classic Amiga environment with your disks attached.
It relies on the filename to select the model of classic Amiga to emulate so if your disk image has the letters AGA in the filename, then an A1200 will be emulated. Otherwise, it falls back to a classic A500 setup. It also supports multi-disk demos and games very simply: If the filenames for your disk images are identical apart from a trailing numerical character, it will “insert” both disks into the system. As an example, here’s what my directory of classic “Grapevine” disk magazines looks like after I’d renamed them according to RunInUAE’s standards:
By default, the emulation runs in fullscreen mode. You can quit the emulation by holding down Ctrl and pressing the Alt and q keys, and Ctrl-Alt-s switches out of fullscreen mode. You can also tweak the settings for an individual game by dropping it onto the RunInUAE window and right-clicking to see a menu where you can change the input devices (joysticks etc.), screenmode and other things. This creates a customised .uaerc file for that particular image which you can see opened in Multiview below:
Here, I’ve switched to windowed mode instead of full-screen. You can also see in the generated .uaerc that it provides a Workbench volume providing a 3.1 installation mounted on a virtual DH0:. This installation lives under System:Emulation/AmigaDisks/Workbench3.1 – this will be important later on… In addition, my usual volumes (Software, Work and System) are also mounted into the virtual Amiga in read-only mode.
A nice touch is that the DiskImageGUI utility also supports Amiga disk images, so you can also mount the images as virtual IDFn: volumes. As an example, I’ve mounted the Grapevine disk image with Mounter and you can see it show up on my X5000 desktop, accessible just like a real floppy image:
The DiskImageGUI runs as a commodity after you close the window, so you can always bring it back with the shortcut Ctrl-Alt-d.
WHDLoad is an awesome tool developed for classic 68k Amigas that makes lots of games and demos runnable straight from a hard drive, even for programs that didn’t natively support a HD install option. There are lots of archives available on the Internet of thousands of classic games and demos, all pre-configured and ready to go. These can also run on the X5000 and AmigaOS 4 through RunInUAE, as it provides a WHDLoad “shim” which boots an emulated Amiga and runs the real WHDLoad from inside that.
However, the bundled WHDLoad binary is pretty old, and is also unregistered so before a game starts you are presented with a “please register” message and a timer that counts down before the game will actually start. Newer versions of WHDLoad include numerous improvements and also remove the registration requirement, so the first thing I did was update the installation. I went to the WHDLoad homepage and downloaded the latest “usr” package which was a standard LHA archive. After extracting it, I ran through the installer utility, but (and this is important!) I selected the C directory under System:Emulation/AmigaDisks/Workbench3.1 as the destination. Do not install this into your AmigaOS 4 System volume, as otherwise it’ll overwrite the WHDLoad “shim”. You can see the correct path specified in the screenshot below:
Once this was done, all my WHDLoad games and demos worked exactly as they did on my A1200! Here’s what the emulation screen displays briefly before the game starts, showing the updated WHDLoad install:
The only AmigaInput supported gamepad I have for my X5000 doesn’t let me use the “hat” switchpad in most games. For some things this is fine, but for other games I prefer the feel of the classic pad instead of using the joystick nubs. I found a fantastic tool on http://os4depot.net/ called AmigaInputAnywhere which lets you use an AmigaInput device to trigger keypresses or other actions. With this, I could set my gamepad’s “hat” controller to trigger the numeric keys (8=Up, 2=Down, 4=Left, 6=Right, 5=Fire) which can then be selected as an emulated Joystick in RunInUAE. For the games where I preferred the Hat switch, I just opened up RunInUAE and dropped the game icons on the open window, right clicked and selected Per-game settings –> Set joystick to –> kbd1 (number pad). In the next screenshot, you can see the AmigaInput prefs open, along with my AmigaInputAnywhere configuration and a test session in the shell where moving the hat switchpad generated a stream of number-pad keypresses:
Full system emulation
Although I could now run my old games and demos through RunInUAE, I also wanted to emulate my A1200 as a “full system” which boots into my Workbench 3.9 installation so I can continue with my old projects. I had the directories prepared with the contents of my A1200’s hard disk copied over; this was covered in my earlier blog post here.
To setup this emulation environment, I decided to use an updated E-UAE which has been released on OS4Depot here. This version of E-UAE has JIT support which can add a significant speed boost, although it’s worth experimenting with it as some things can become unstable when enabling it. I also definitely didn’t want to interfere with my stock E-UAE installation. Even though it’s old, it still seems to work fine for my disk images and WHDLoad games, so instead I unpacked this updated release into a seperate directory at Software:Programs/E-UAE_1.0.0.
I then created the Configuration file at Software:Programs/E-UAE_1.0.0/.uaerc based on a template from Amigaone X5000 Blog. I made a few simple changes: Disabled the floppy drive sound, enabled the JIT (after some experimentation, my installation seemed to run better with it) and changed the paths to my emulated hard drives and ROM image. For reference, my slightly modified .uaerc file is listed below:
I created a little script file to launch the emulation environment, this was a simple text file with the following lines:
I saved this as “a1200” in my Work:c directory I keep for my own scripts and tools (this is also added to my path in my S:User-Startup). I also ran Protect Work:c/a1200 +s to mark it as an executable script. So that I could also launch this from X-Dock, I copied over a nice .info icon file to this directory and renamed it as a1200.info. I then modified the tooltypes as per the screenshot below:
This sets the default tool to C:IconX which launches script files, and also set the window to the NIL: device so I didn’t get a screen of UAE debugging output.
With that in place, all I have to do now is click on the icon in my X-Dock, and I get a windowed emulation of my A1200 just as I’d left it!
I can easily put this into fullscreen mode by using the Ctrl-Alt-s shortcut, and I also discovered that Ctrl-Alt-g “grabs” the mouse pointer inside the emulation which is really useful when it’s running in windowed mode.
Classic 68k binaries
AmigaOS 4 also includes a built-in JIT 68k emulation layer called “Petunia” that allows classic system-legal 68k binaries and libraries to run unmodified. For these to run, they must only use OS system calls and avoid “hitting the hardware” – so anything that bangs on the custom chipset can’t be used (although there’s always RunInUAE for such things).
Many tools and utilities work just fine however, especially more recent 68k binaries. In the screenshot below, I’m running the 68k binaries of the Workbench 3.1 “calculator” tool, and also AmTelnet for some classic BBS action. AmTelnet is a particularly good example of what can run: It’s a full MUI program which also installs some custom 68k libraries. Apart from the occasional Grim Reaper crash on launch, it functions just like a PPC-native program. It’s fully integrated into the updated Workbench environment, and if I never looked at the task listing in SysMon and saw the “68k” field in the “Arch” column, I’d have thought it was a native application!
So there you have it! I hope that’s been of some help and interest :) I’ll continue to update this blog with more of my Amiga experiments, and feel free to leave a comment with any questions!