Unix Archive
The contents of this repository allow older versions of UNIX (ancient UNIX) to run easily on modern Unix-like systems (Linux, FreeBSD, macOS, among others). ↫ Run ancient UNIX GitHub page With the guides in this repository, you can easily run Versions 1/5/7 UNIX and 2.11BSD UNIX for the PDP-11 and Version 7 UNIX for x86 (ported to x86 by Robert Nordier in 1999, with patches in 2006-2007). That’s it.
Jonathan Pallant got lucky and managed to score a massive haul of ’90s UNIX workstations, one of which was an HP 9000 Model 340, a HP-UX workstation built around a Motorola 68030 processor at 16.7 MHz. It doesn’t come with a hard drive or even a floppy controller, though, so he decided to borrow a PA-RISC-based HP 9000 Model 705 to set up an HP-UX 9 cluster. But wait, how does that work, when we’re dealing with two entirely different architectures? What’s more fun though, is putting it into a cluster with the Model 705 and network booting it. Yes, that a 68030 machine network booting from a PA-RISC machine … and sharing the same root filesystem. But aren’t PA-RISC binaries and 68K binaries quite different? Oh yes, they really are. So, how does that work? ↫ Jonathan Pallant HP-UX is far more interesting and fascinating than a lot of people give it credit for, and while my interest lies with HP-UX 11i, I find what Pallant is doing here with HP-UX 9 just as fascinating. You first need to install HP-UX 9 for PA-RISC on the 700 series machine, convert it to a cluster server, and then install HP-UX 9 for 68k on top of that PA-RISC installation. After this is done, you effectively end up with a single root file system that contains both PA-RISC and 68k binaries, and you can network boot the 68k-based Model 340 right from it – using the same root filesystem on both machines. Absolutely wild. No, these are not universal binaries or some other trick you might know of from more modern system. In fact, installing the 68k version of HP-UX 9 “into” the PA-RISC HP-UX 9 cluster server, you end up with something called a Context Dependent Filesystem. To get a better idea of what this means and how this works, you should really head on over to Pallant’s excellent article for all the details.
A unique and very important find at the University of Utah: while cleaning out some storage rooms, the staff at the university discovered a tape containing a copy of UNIX v4 from Bell Labs. At this time, no complete copies are known to exist, and as such, this could be a crucial find for the archaeology of early UNIX. The tape in question will be sent to the Computer History Museum for further handling, where bitsavers.org will conduct the recovery process. I have the equipment. It is a 3M tape so it will probably be fine. It will be digitized on my analog recovery set up and I’ll use Len Shustek’s readtape program to recover the data. The only issue right now is my workflow isn’t a “while you wait” thing, so I need to pull all the pieces into one physical location and test everything before I tell Penny it’s OK to come out. ↫ bitsavers.org It’s amazing how we still manage to find such treasures in nooks and crannies all over the world, and with everything looking good so far, it seems we’ll soon be able to fill in more of UNIX’ early history.
This is a conversion of the original V7 pwd program for use on POSIX systems (tested primarily on Linux). This is mostly of historical interest — modern systems have a library routine or system call for getting the current directory, and don’t need this. I’ve attempted to make the minimum set of logic/functionality changes needed to make the program work, preserving the core of the original logic. I’ve made slightly more aesthetic changes, to make reading easier for a post-standardization C speaker. ↫ Cliff L. Biffle Over on Fedi, Cliff L. Biffle provides more details as to why he undertook this project.
Chris Siebenmann with another interesting look at a tiny detail of UNIX history. A few years ago I wrote about the divide in chown() about who got to give away files, where BSD and V7 were on one side, restricting it to root, while System III and System V were on the other, allowing the owner to give them away too. The answer is that the restriction was added in V6, where the V6 chown(2) manual page has the same wording as V7. In Research Unix V5 and earlier, people can chown(2) away their own files; this is documented in the V4 chown(2) manual page and is what the V5 kernel code for chown() does. This behavior runs all the way back to the V1 chown() manual page, with an extra restriction that you can’t chown() setuid files. ↫ Chris Siebenmann The deeper levels of this particular rabbit hole need more exploring, though, as eventually Siebenmann hits a roadblock when trying to figure out why, exactly, the restriction was added, and why certain versions chose to not adopt the new restriction. This may be part of the lore of UNIX we won’t uncover, until one of the people involved speaks up.
I’ve been working on developing an operating system for the TI-99 for the last 18 months or so. I didn’t intend this—my original plan was to develop enough of the standard C libraries to help with writing cartridge-based and EA5 programs. But that trek led me quickly towards developing an OS. As Unix is by far my preferred OS, this OS is an approximation. Developing an OS within the resources available, particularly the RAM, has been challenging, but also surprisingly doable. ↫ UNIX99 forum announcement post We’re looking at a quite capable UNIX for the TI-99, with support for its sound, speech, sprites, and legacy 9918A display modes, GPU-accelerated scrolling, stdio (for text and binary files) and stdin/out/err support, a shell (of course), multiple user support, cooperative tasks support, and a ton more. And remember – all of this is running on a machine with a 16-bit processor running at 16MHz and a mere 16KB of RAM. Absolutely wild.
It may be arcane knowledge to most users of UNIX-like systems today, but there is supposed to be a difference between /usr/bin and /usr/sbin; the latter is supposed to be for “system binaries”, not needed by most normal users. The Filesystem Hierarchy Standard states that sbin directories are intended to contain “utilities used for system administration (and other root-only commands)”, which is quite vague when you think about it. This has led to UNIX-like systems basically just winging it, making the distinction almost entirely arbitrary. For a long time, there has been no strong organizing principle to /usr/sbin that would draw a hard line and create a situation where people could safely leave it out of their $PATH. We could have had a principle of, for example, “programs that don’t work unless run by root”, but no such principle was ever followed for very long (if at all). Instead programs were more or less shoved in /usr/sbin if developers thought they were relatively unlikely to be used by normal people. But ‘relatively unlikely’ is not ‘never’, and shortly after people got told to ‘run traceroute’ and got ‘command not found’ when they tried, /usr/sbin (probably) started appearing in $PATH. ↫ Chris Siebenmann As such, Fedora 42 unifies /usr/bin and /usr/sbin, which is kind of a follow-up to the /usr merge, and serves as a further simplification and clean-up of the file system layout by removing divisions and directories that used to make sense, but no longer really do. Decisions like these have a tendency to upset a small but very vocal group of people, people who often do not even use the distribution implementing the decisions in question in the first place. My suggestions to those people would be to stick to distributions that more closely resemble classic UNIX. Or use a real UNIX. Anyway, these are good moves, and I’m glad most prominent Linux distributions are not married to decisions made in the ’70s, especially not when they can be undone without users really noticing anything.
Earlier this year, I was trying to get actual daily work done on HP-UX 11.11 (11i v1) running on HP’s last and greatest PA-RISC workstation, the HP c8000. After weeks of frustration caused first by outdated software no longer working properly with the modern web, and then by modern software no longer compiling on HP-UX 11.11, I decided to play the ace up my sleeve: NetBSD’s pkgsrc has support for HP-UX. Sadly, HP-UX is obviously not a main platform or even a point of interest for pkgsrc developers – as it should be, nobody uses this combination – so various incompatibilities and more modern requirements had snuck into pkgsrc, and I couldn’t get it to bootstrap. I made some minor progress here and there with the help of people far smarter than I, but in the end I just lacked the skills to progress any further. This story will make it to OSNews in a more complete form, I promise. Anyway, in May of this year, it seems Brian Robert Callahan was working on a very similar problem: getting pkgsrc to work properly on IBM’s AIX. The state of packages on AIX genuinely surprises me. IBM hosts a repository of open source software for AIX. But it seems pretty sparse compared to what you could get with pkgsrc. Another website offering AIX packages seems quite old. I think pkgsrc would be a great way to bring modern packages to AIX. I am not the first to think this. There are AIX 7.2 pkgsrc packages available at this repository, however all the packages are compiled as 32-bit RISC System/6000 objects. I would greatly prefer to have everything be 64-bit XCOFF objects, as we could do more with 64-bit programs. There also aren’t too many packages in that repository, so I think starting fresh is in our best interest. As we shall see, this was not as straightforward as I would have hoped. ↫ Brian Robert Callahan Reading through his journey getting pkgsrc to work properly on AIX, I can’t help but feel a bit better about myself not being to get it to work on HP-UX 11.11. Callahan was working with AIX 7.2 TL4, which was released in November 2019 and still actively supported by IBM on a maintained architecture, while I was working with HP-UX 11.11 (or 11i v1), which last got some updates in and around 2005, running on an architecture that’s well dead and buried. Looking at what Callahan still had to figure out and do, it’s not surprising someone with my lack of skill in this area couldn’t get it working. I’m still hoping someone far smarter than I stumbles upon a HP c8000 and dives into getting pkgsrc to work on HP-UX, because I feel pkgsrc could turn an otherwise incredibly powerful HP c8000 from a strictly retro machine into something borderline usable in the modern world. HP-UX is much harder to virtualise – if it’s even possible at all – so real hardware is probably going to be required. The NetBSD people on Mastodon suggested I could possibly give remote access to my machine so someone could dive into this, which is something I’ll keep under consideration.
The most unlikely subsystem of contention is definitely the init system used by Linux, with most popular distributions opting for systemd, while a vocal minority prefers to use something else. Neither of these two groups are wrong or right, as we live in a free world and different people have different needs and desires. Personally, I don’t think there’s a more utterly pointless and meaningless debate than this, and people who make the init system they use their entire personality more often than not come across as really, really sad. It’s a tool; use the one you like and move on with life. A brand new init system was recently released by Leah Neukirchen, who among a ton of other things, contributes to Void Linux. It’s called nitro, and it’s a “tiny process supervisor that also can be used as pid 1 on Linux”, and it also can be used on FreeBSD supervised by FreeBSD’s init. There’s some overlap with runit here, so Neukirchen published a blog post detailing the differences between the two, which should help in getting a better understanding of what makes nitro stand apart. While both use a directory of services managed by small scripts, nitro seems to opt for a more contained, monolithic approach, as it keeps everything in a single process. On top of that, Nitro contains some new features runit doesn’t have. The focus seems to be on integrating a few capabilities that on runit require hacks, but on nitro are just built-in, like “support for one-shot ‘services’, i.e. running scripts on up/down without a process to supervise (e.g. persist audio volume, keep RNG state)”, running service directories multiple times, and more. Nitro also maintains its runtime state in RAM and provides an IPC service to query it, meaning it can be run on read-only filesystems without special configuration. There’s a lot more information in Neukirchen’s blog post, including a look at some of the current limitations of Nitro. I highly suggest reading it, and perhaps we will see nitro as another valid alternative to the popular systemd.
The first prototype was ready in just six months. By October 1986, the project was announced, and in January 1987, the first NEWS workstation, the NWS 800 series, officially launched. It ran 4.2BSD UNIX and featured a Motorola 68020 CPU. Its performance rivaled that of traditional super minicomputers, but with a dramatically lower price point ranging from ¥950,000 to ¥2.75 million (approximately $6,555 to $18,975 USD in 1987). Competing UNIX workstations typically cost closer to ¥10 million (around $69,000 USD). NEWS caught on quickly in universities and R&D labs, where cost sensitive researchers needed real performance. The venture team had invested ¥400 million into development (about $2.76 million USD), and remarkably, they recouped those costs within just two months of launch. That same year, Sony introduced a lower cost version called POP NEWS (PWS 1550). With a GUI shell named NEWS Desk, a document sharing format called CDFF (Common Document File Format), and a focus on Japanese language desktop publishing, PopNEWS aimed to make UNIX more accessible to general business users. Targeted at the Desktop Publishing market, it showed Sony’s desire to bridge consumer and professional segments in ways no other UNIX vendor was trying at the time. ↫ Obsolete Sony’s Newsletter I’ve been fascinated by Sony’s NEWS workstations, and especially the NEWS-OS operating system, for a long time now. Real hardware is hard to find and prohibitively expensive, but some of these Sony NEWS workstations can be emulated through MAME. Sadly, as far as I can tell, you can only emulate NEWS-OS up to version 4.x, as I haven’t been able to find any information about emulating version 5.x and the final version, 6.x. If anyone knows anything about how to emulate these, if at all possible, please do share with the rest of us. What’s interesting about Sony’s UNIX workstation efforts from the ’80s and ’90s is that they played an important role in the early development of the PlayStation. The early development kits for the PlayStation were modified NEWS workstations, with added PlayStation hardware. To further add to the importance of the NEWS line for gaming, Nintendo used them to develop several influential and popular first-party SNES titles, which isn’t surprising considering Nintendo and Sony originally worked together on bringing a CD-ROM drive to the SNES, which would later morph into the PlayStation as Nintendo cancelled the agreement at the last second.
If you use Unix today, you can enjoy relatively long file names on more or less any filesystem that you care to name. But it wasn’t always this way. Research V7 had 14-byte filenames, and the System III/System V lineage continued this restriction until it merged with BSD Unix, which had significantly increased this limit as part of moving to a new filesystem (initially called the ‘Fast File System’, for good reasons). You might wonder where this unusual number came from, and for that matter, what the file name limit was on very early Unixes (it was 8 bytes, which surprised me; I vaguely assumed that it had been 14 from the start). ↫ Chris Siebenmann I love these historical explanations for seemingly arbitrary limitations.
There’s a number of backups of old DECtapes from Dennis Ritchie, which he gave to Warren Toomey in 1997. The tapes were eventually uploaded, and through analysis performed by Yufeng Gao, a lot of additional details, code, and software were recovered from them. A few days ago, Gao came back with the results from their analys of two more tapes, and on it, they found something quite special. Getting this recovered version to run was a bit of a challenge, and only aap’s PDP-11 emulator is capable of running it. To even get it to run in the first place, Gao had perform quite some intricate steps, but eventually he managed to build an image that can be downloaded and booted on aap’s PDP-11 emulator. The image in question, as well as some more details, can be found on the GitHub page.
What might be somewhat more surprising though considering its research origins is that Unix almost since the very beginning had a comprehensive set of online reference documentation for all its commands, system calls, file formats, etc. These are the the manual- or man-pages. On Unix systems used interactively, the man-pages have historically always been installed, space permitting. The way the manual pages have evolved and how they are used has changed over the decades. This set of posts is intended to give people unfamiliar with them an overview, as well as offer a review to seasoned users. ↫ Alex Bochannek Right in this first article in the series there’s an interesting observation I never stopped and thought about: because the original creators of UNIX were writing the content of man pages with the very tools they were creating for UNIX, it led to a virtuous cycle. “Unix tools were used to document Unix, improving the documentation tools themselves as well.” I tend to use the internet now to learn how specific tools and commands work, but having such detailed man pages built right into the operating system was a huge deal pre-internet.
It kind of goes by under the radar, but aside from HP-UX, Solaris, and AIX, there’s another traditional classic UNIX still in active development today: UnixWare (and its sibling, OpenServer). Owned and developed by Xinuos, UnixWare and other related code and IP was acquired by them when the much-hated SCO crashed and burned about 15 years ago or so, and they’ve been maintaining it ever since. About a year ago, Xinuos released Update Pack 1 and Maintenance Pack 1 for UnixWare 7 Definitive 2018, followed by similar update packs for OpenServer 6 later in 2024. These update packs bring a bunch of bugfixes and performance improvements, as well as a slew of updated open source components, like new versions of SAMBA, sendmail, GCC and tons of other GNU components, OpenSSH and OpenSSL, and so, so much more, enabling a relatively modern and up-to-date build and porting environment. They can be installed through the patchck update utility, and while the Maintenance Pack is free for existing registered users, the Update Pack requires a separate license. UnixWare, while fully capable as a classic UNIX for workstations, isn’t really aimed at individuals or hobbyists (sadly), and instead focuses on existing enterprise deployments, where such licensing costs are par for the course. UnixWare runs on x86, and can be installed both on real hardware as well as in various virtualised environments. I contacted Xinuos a few days ago for a review license, and they supplied me with one so I can experiment with and write about UnixWare. I’ve currently got it installed in a Linux kvm, where it runs quite well, including the full X11R6 CDE desktop environment and graphical administration tools. Installing updates is a breeze thanks to patchck automating the process of finding, downloading, and installing the correct ones. I intend to ask Xinuos about an optimal configuration for running UnixWare on real hardware, too.
Update: there’s a fork called heirloom-ng that is actually still somewhat maintained and contains some more changes and modernisations compared to the old version. The Heirloom Project provides traditional implementations of standard Unix utilities. In many cases, they have been derived from original Unix material released as Open Source by Caldera and Sun. Interfaces follow traditional practice; they remain generally compatible with System V, although extensions that have become common use over the course of time are sometimes provided. Most utilities are also included in a variant that aims at POSIX conformance. On the interior, technologies for the twenty-first century such as the UTF-8 character encoding or OpenType fonts are supported. ↫ The Heirloom Project website I had never heard of this before, but I like the approach they’re taking. This isn’t just taking System V tools and making them work on a modern UNIX-like system as-is; they’re also improving by them adding support for modern technologies, without actually changing their classic nature and the way old-fashioned users expect them to work. Sadly, the project seems to be dead, as the code hasn’t been altered since 2008. Perhaps someone new is willing to take up this project? As it currently stands, the tools are available for Linux, Solaris, Open UNIX, HP-UX, AIX, FreeBSD, NetBSD, and OpenBSD, but considering how long the code hasn’t been touched, I wonder if they still run and work on any of those systems today. They also come in various different versions which comply with different variants of the POSIX standard.
As an online discussion grows longer, the probability of a someone mentioning macOS is a UNIX approaches 1. In fact, it was only late last year that The Open Group announced that macOS 15.0 was, once again, certified as UNIX, continuing Apple’s long-standing tradition of certifying macOS releases as “real” UNIX®. What does any of this actually, mean, though? Well, it turns out that if you actually dive into Apple’s conformance statements for macOS’ UNIX certification, it doesn’t really mean anything at all. First and foremost, we have to understand what UNIX certification really means. In order to be allowed to use the UNIX trademark, your operating system needs to comply with the Single UNIX Specification (SUS), which specifies programming interfaces for C, a command-line shell, and user commands, more or less identical to POSIX, as well as the X/Open Curses specification. The latest version is SUS version 4, originally published in 2008, with amendments published in 2013 and 2016, which were rolled up into version 4 in 2018. The various versions of the SUS that exist, in turn, correspond to a specific UNIX trademark. In table form: Trademark SUS version SUS published in: SUS last amended in: UNIX® 93 n.a. n.a. n.a. UNIX® 95 Version 1 1994 n.a. UNIX® 98 Version 2 1997 n.a. UNIX® 03 Version 3 2002 2004 UNIX® V7 Version 4 2008 2016 (2018 for roll-up) When you read that macOS is a certified UNIX, which of these versions and trademarks do you assume macOS complies with? You’d assume they would just target the latest trademark and SUS version, right? This would allow macOS to carry the UNIX® V7 trademark, because they would conform to version 4 of the SUS, which dates to 2016. The real answer is that macOS 15.0 only conforms to version 3 of the SUS, which dates all the way back to the ancient times of 2004, and as such, macOS is only UNIX® 03 (on both Intel and ARM). However, you can argue this is just semantics, since it’s not like UNIX and POSIX are very inclined to change. So now, like the UNIX nerd that you are, you want to see all this for yourself. You use macOS, safe in the knowledge that unlike those peasants using Linux or one of the BSDs, you’re using a real UNIX®. So you can just download all the tests suites (if you can afford them, but that’s a whole different can of worms) and run them, replicating Apple’s compliance testing, seeing for yourself, on your own macOS 15 installation, that macOS 15 is a real UNIX®, right? Well, no, you can’t, because the version of macOS 15 Apple certifies is not the version that’s running on everyone’s supported Macs. To gain its much-vaunted UNIX certification for macOS, Apple cheats. A lot. The various documents Apple needs to submit to The Open Group as part of the UNIX certification process are freely available, and mostly it’s a lot of very technical questions about various very specific aspects of macOS’ UNIX and POSIX compliance few of us would be able to corroborate without extensive research and in-depth knowledge of macOS, UNIX, and POSIX. However, at the end of every one of these Conformance Statements, there’s a text field where the applicant can write down “additional, explanatory material that was provided by the vendor”, and it’s in these appendices where we can see just how much Apple has to cheat to ensure macOS passes the various UNIX® 03 certification tests. In the first of these four documents, Internationalised System Calls and Libraries Extended V3, Apple’s “additional, explanatory material” reads as follows: Question 27: By default, core file generation is not enabled. To enable core file generation, you can issue this command: sudo launchctl limit core unlimited Testing Environment Addendum: macOS version 15.0 Sequoia, like previous versions, includes an additional security mechanism known as System Integrity Protection (SIP). This security policy applies to every running process, including privileged code and code that runs out of the sandbox. The policy extends additional protections to components on disk and at run-time, only allowing system binaries to be modified by the system installer and software updates. Code injection and runtime attachments to system binaries are no longer permitted. To run the VSX conformance test suite we first disable SIP as follows: – Shut down the system.– Press and hold the power button. Keep holding it while you see the Apple logo and the message “Continue holding for startup options”– Release the power button when you see “Loading startup options”– Choose “Options” and click “Continue”– Select an administrator account and enter its password.– From the Utilities menu in the Menu Bar, select Terminal.– At the prompt, issue the following command: “csrutil disable”– You should see a message that SIP is disabled. From the Apple menu, select “Restart”. By default, macOS coalesces timeouts that are scheduled to occur within 5 seconds of each other. This can randomly cause some sleep calls to sleep for different times than requested (which affects tests of file access times) so we disable this coalescing when testing. To disable timeout coalescing issue this command: sudo sysctl -w kern.timer.coalescing_enabled=0 By default there is no root user. We enable the root user for testing using the following series of steps:– Launch the Directory Utility by pressing Command and Space, and then typing “Directory Utility”– Click the Lock icon in Directory Utility and authenticate by entering an Administrator username and password.– From the Menu Bar in Directory Utility:– Choose Edit -> Enable Root User. Then enter a password for the root user, and confirm it.– Note: If you choose, you can later Disable Root User via the same menu. ↫ Apple’s appendix to Internationalised System Calls and Libraries Extended V3 The second conformance statement, Commands and Utilities V4, has another appendix, and it’s a real doozy (the indicate repeat remarks from the previous appendix; I’ve removed them for brevity): Testing Environment Addendum: The third and fourth conformance statements have
How do you fit a 250kB dictionary in 64kB of RAM and still perform fast lookups? For reference, even with modern compression techniques like gzip -9, you can’t compress this file below 85kB. In the 1970s, Douglas McIlroy faced this exact challenge while implementing the spell checker for Unix at AT&T. The constraints of the PDP-11 computer meant the entire dictionary needed to fit in just 64kB of RAM. A seemingly impossible task. ↫ Abhinav Upadhyay They still managed to do it, but had to employ some incredibly clever tricks to make it work, and make it work fast. Such skillful engineers interested in optimising and eeking the most possible performance out of underpowered hardware still exist today, but they’re not in any position to make lasting changes at any of the companies defining our technology today. Why spend money on skilled engineers, when you can just throw cheap hardware at the problem? I wonder just how many resources the spellchecking feature in Word or LibreOffice Writer takes up.
One of the innovations that the V7 Bourne shell introduced was built in shell wildcard globbing, which is to say expanding things like *, ?, and so on. Of course Unix had shell wildcards well before V7, but in V6 and earlier, the shell didn’t implement globbing itself; instead this was delegated to an external program, /etc/glob (this affects things like looking into the history of Unix shell wildcards, because you have to know to look at the glob source, not the shell). ↫ Chris Siebenmann I never knew expanding wildcars in UNIX shells was once done by a separate program, but if you stop and think about the original UNIX philosophy, it kind of makes sense. On a slightly related note, I’m currently very deep into setting up, playing with, and actively using HP-UX 11i v1 on the HP c8000 I was able to buy thanks to countless donations from you all, OSNews readers, and one of the things I want to get working is email in dtmail, the CDE email program. However, dtmail is old, and wants you to do email the UNIX way: instead of dtmail retrieving and sending email itself, it expects other programs to those tasks for you. In other words, to setup and use dtmail (instead of relying on a 2010 port of Thunderbird), I’ll have to learn how to set up things like sendmail, fetchmail, or alternatives to those tools. Those programs will in turn dump the emails in the maildir format for dtmail to work with. Configuring these tools could very well be above my paygrade, but I’ll do my best to try and get it working – I think it’s more authentic to use something like dtmail than a random Thunderbird port. In any event, this, too, feels very UNIX-y, much like delegating wildcard expansion to a separate program. What this also shows is that the “UNIX philosophy” was subject to erosion from the very beginning, and really isn’t a modern phenomenon like many people seem to imply. I doubt many of the people complaining about the demise of the UNIX philosophy today even knew wildcard expansion used to be done by a separate program.
While we’re out here raising funds to make me daily-drive HP-UX 11i v1 – we’re at 59% of the goal, so I’m starting to prepare for the pain – it seems you can actually run older versions, HP-UX 10.20 and 11.00 to be specific, in a virtual machine using QEMU. QEMU is an open source computer emulation and virtualization software, first released in 2003 by Fabrice Bellard. It supports many different computer systems and includes support for many RISC architectures besides x86. PA-RISC emulation has been included in QEMU since 2018. QEMU emulates a complete computer in software without the need for specific virtualization hardware. With QEMU, a full HP Visualize B160L and C3700 workstation can be emulated to run PA-RISC operating systems like HP-UX Unix and compatible applications. ↫ Paul Weissman at OpenPA The emulation is complete enough that it can run X11 and CDE, and you can choose between emulating 32bit PA-RISC of 64bit PA-RISC. Devices and peripherals support is a bit of a mixed bag, with things like USB being only partially supported, and audio not working at all since an audio chip commonly found in PA-RISC workstations isn’t supported either. A number of SCSCI and networking devices found on HP’s workstations aren’t supported either, and a few chipsets don’t work either. As far as operating system support goes, you can run HP-UX 10.20, HP-UX 11.00, Linux, and NetBSD. Newer (11i v1 and later) and older (9.07 and 9.05) versions of HP-UX don’t work, and neither does NeXTSTEP 3.3. Some of these issues probably stem from missing QEMU drivers, others from a lack of testing; PA-RISC is, after all, not one of the most popular of the dead UNIX architectures, with things like SPARC and MIPS usually taking most of the spotlight. Absolutely nothing beats running operating systems on the bare metal they’re designed for, but with PA-RISC hardware becoming ever harder to obtain, it makes sense for emulation efforts to pick up speed so more people can explore HP-UX. I’m weirdly into HP-UX, despite its reputation as a difficult platform to work with, so I personally really want actual hardware, but for most of you, getting HP-UX 11i to work properly on QEMU is most likely the only way you will ever experience this commercial UNIX.
I commented on Lobsters that /tmp is usually a bad idea, which caused some surprise. I suppose /tmp security bugs were common in the 1990s when I was learning Unix, but they are pretty rare now so I can see why less grizzled hackers might not be familiar with the problems. I guess that’s some kind of success, but sadly the fixes have left behind a lot of scar tissue because they didn’t address the underlying problem: /tmp should not exist. ↫ Tony Finch Not only is this an excellent, cohesive, and convincing argument against the existence of /tmp, it also contains some nice historical context as to why things are the way they are. Even without the arguments against /tmp, though, it just seems entirely more logical, cleaner, and sensible to have /tmp directories per user in per user locations. While I never would’ve been able to so eloquently explain the problem as Finch does, it just feels wrong to have every user resort to the exact same directory for temporary files, like a complex confluence of bad decisions you just know is going to cause problems, even if you don’t quite understand the intricate interplay.
