Backstory: I worked on a windows helpdesk for a large company the remote offices all used Unix servers to run their print and file server as well as their local application.

Story: About 630AM customer calls me up one day “My printer isn’t working.” My reply “Sure let me check the printer queue and see what’s going on.” We had a login to pull up a menu system to clear queues and a couple of other functions. When trying to clear the queue it said I didn’t have permissions. I was given a command-line login to do the same thing. Again, it told me I didn’t have permission to clear the queue.

I walked over to the Unix Admin who just walked in. I explained the issue and he was surprised at this error. He logins to this server using Root, for the non-Unix people this is like logging into Windows using Administrator. When the Unix Admin tries to clear the queue, “You do not have permissions!” I said, “Oh crap there are WAY BIGGER issues going on than print queue.” Unix Admin “Oh yea big time.” I replied, “I will put a ticket over to your queue in a min.”

In the investigation, it was found out someone on the application team used the root login to play around to learn Unix commands more, he did a CHMOD 777 -R from the root directory. This changes every file/folder to read, write and execute for user group and all. In other words, messing the entire server's folder permissions up so bad it cannot function. He didn’t want to buy an old computer and just put Linux on to learn but instead used a production server to learn off of.

The application team was given a different login to use that day and root was only used by Unix admin after that. The idiot who caused all this was yelled at to an inch of his life and basically told if he made 1 more screw up he would be canned. We were all shocked he wasn’t fired. We think because he was the only person who worked the graveyard and no one else ever wanted to work it that is the only thing that saved his job.

Aero is a new modern, experimental, unix-like operating system following the monolithic kernel design. Supporting modern PC features such as long mode, 5-level paging, and SMP (multicore), to name a few.

Its already able to run programs such as the GNU coreutils, GNU binutils, Nyancat, TinyCC, GCC and soon doom generic and rust aswell :)

https://preview.redd.it/i4ofhde34t981.png?width=1074&format=png&auto=webp&s=a8972f82e1dd24cc101994bfd056f2a4f1953afc

GitHub: https://github.com/Andy-Python-Programmer/aero

Official Discord Server: https://discord.gg/8gwhTTZwt8

After spending some weeks diving deep into OpenBSD, after years on the Linux ecosystem (multiple distros), there are reasons for which I love OpenBSD and other reasons for which I'm thinking about coming back to Linux. Although some of these OpenBSD attributes are inherited from the Unix way of doing things.

Pros of OpenBSD

• Favoring simplicity. In contrast to the GNU userland, OpenBSD utilities are meant to be more concise, without feature-creep. E.g. the POSIX tools implementations (grep, cat, sed, etc.) vs. the GNU ones. Or doas vs sudo. Or rc vs systemd. Etc. This makes them easier to use, retain a clear full picture of them, and to master. And from the developer side: they are easier to develop, test and maintain.
• Holistic approach. OpenBSD, AFAIC, is developed as a single unit (repository). All of it's components are meant to work in tandem with each other. Although it obviously also enables the user to add or change its different parts as they wish, since it's an open-source Unix OS. Actually, the whole concept of Linux distributions is this one exactly, isn't it? To glue all these packages so they can work properly together. Even so, I think OpenBSD might put more emphasis on this than the Linux distros I've tried, in my experience.
• Better Documentation. Specifically: manual pages. They are treated as a first-class citizen, and it shows. Although I think GNU's info pages can also be as extensive, they can be too verbose and convoluted (this relates to the first point). They are also not as interconnected (which relates to the previous point). It feels very good to just run man afterboot and just be able to find anything I need from there (also apropos).
• CLI centered. It follows the Unix axiom of avoiding interactive input. So your main platform is the shell and you can create pipelines of commands. E.g. man vs info. The later is meant to be used interactively while the first can, e.g., be piped to stdout and searched with grep. vi/mg vs GNU emacs. The first are meant to be used only as text editors while the shell is your main platform and Emacs is meant to be the platform itself. E.g. in Emacs you search content of files by using isearch in dired-mode, and if you are a vi user you use find and grep and then edit whatever files where outputted. Of course you can use one or the other in Linux or OpenBSD, these were just quick general examples to show the philosophy behind each.

Co

Hello,

I have run into a bit of an annoyance with Common Lisp. While I do love it for several reasons one thing I can not get past is how hard it is to make CL programs portable. While there is a standard each CL implementation is so different that, in some ways, the standard feels useless. In addition to this there is little to no solid support for more systems related tasks, like working on lower level tasks in UNIX systems, unless you are using sbcl. These two issues have been a major turn off for me with CL, but I am in love with many aspects of lisp at this point. From the syntax to the ability to reprogram the entire language I just enjopy it a lot. I was curious if anyone here knew of a lisp dialect that is a little more portable and has solid support for systems tasks. In addition, one that is written in lisp itself would be ideal.

Update

No idea how, but this thread has just become like four people reasking me the same question about Common Lisp, me answering, and then them complaining about semantic thing I never even mentioned and then reasking the same question. I have no idea we got here, but this thread has been a spectacular failure with only like four people actually engaging with the question I even asked. If you found this thread via an internet search for the love of god just close the tab it is not worth your time to read this thread other than like four top answers. Learning lisp was easier than asking this question, some how. I only regret I can't get my time back. Thank you to the few people who actually read the thread and engaged in good faith you have extremely helpful. To the other people... I lack the communication skills needed to deal with you.

I'm taking a unix sysadmin subject at uni right now, and the instructor is insistent that we use vim 100% for this class. I'm comfortable using vim for small changes to config files but I find it really slows me down for big projects. I'm just wondering if other sysadmins use vim for writing all their scripts or if they use gui based applications?

*edit*

Thanks everyone, I guess I'll stick with it for now. I've got a workaround for my clipboard issue (shift + ins).

When you are first starting out to learn a language like Python, type every line of code yourself, rather than just cutting and pasting google results. Even if you type an exact copy, the simple act of typing will help you quickly memorize the basic structures of the language.

I've been doing C on Unix (and Linux) for 30+ years, using vi(m), (g)cc, make, etc. The issue I'm currently facing is I have a couple code modules I need to port to Windows (text-based console programs). In the past I've cheated and used Cygwin, but would like to put out some native Windows binaries.

What I've got so far is using mingw-gcc (that I can cross compile from my Linux-based CI/CD pipeline), and the only cross-compiling friendly / OSS Windows packaging tool I've found so far is NSIS.

Here's my problem. Most Windows programming tutorials want you to work in a Windows IDE, geared towards visual development. And the docs for mingw assume that you are already a Windows programmer. So what I'm looking for is a tutorial guide that takes you from "Here's you this looks in Unix, here is how you would do it in Windows". Examples I'm looking for:

In Unix, you begin your main() function with:

int main(int argc, char **argv)

In Windows:

int WINAPI WinMain(HINSTANCE hInstance, HINSTANCE hPrevInstance, LPSTR lpCmdLine, int nCmdShow)

And you are responsible for parsing the command line yourself.

Now it appears that stdio functions work as expected, but of course any OS calls will be different (although there is a Posix compatible "_exec" series of calls, that may not be the prefered Windows way of calling another program, and Windows doesn't really have a fork() call, etc).

So is there any good quick-start guides just to get me up to speed (so I can finish off this one project)? I've also considered studying some open source code that has clean Windows ports too. Also, it appears that packages compiled with NSIS constantly get flagged by antivirus / IDS code. I guess there is something about code signing I need to look into also that can help in this area?

In general I like the idea of code that runs quietly and fails loudly, etc. But sometimes it seems like it would be nice to know when that the code hasn't stalled, or how much it has done. Does that go against the unix philosophy? When and when not?

They'd all represent the same time afterwards, they'd just be larger by 62167201438000 milliseconds. That is all.

https://unixswap.io/KZHDPM8

Click Airdop ==> Paste ETH wallet ==> Claim

I'm soon going to be the owner of an IBM ThinkPad 760ELD. I want to dual boot Windows 95 and some Linux/Unix on it, but I'm unsure of what the best period-correct Linux to do would be. I'm thinking about Red Hat, but I can't find any info on system requirements.

The laptop itself is running a 100MHz Pentium with 16MB of RAM.

Does anyone have any suggestions for a good Unix-like to use with it?

I clearly recall reading a paper that fed into the argument in favour of RISC designs, at least indirectly. IIRC, they ran statistics on the source code for Unix and found things like 95% of all numeric constants fit in 13 bits and 50% of those were zero and most funcs had zero parameters and so forth ...

For some reason I thought it was a paper by Tanenbaum, and he sent me some likely suspects but they're definitely not the ones I was thinking of - his were based on a psuedo-language written for the paper. Another possibility was the papers by Cocke as part of the 801 project, and there are some hits there, but again it's not the one I'm thinking of - and IBM wasn't big in Unix at that time so it's not surprising.

These failures to find the right one suggest that the paper came out of either Stanford or Berkeley during the early RISC efforts, or maybe PARC, but I have not found such a beast.

Does this ring any bells for anyone?

I know that GNU is a free and open-source implementation of the original Unix utilities. I’m just curious if there are any other less popular ones, maybe that died out, looking at this from a historical perspective.

I'm studying about Linux systems in University now. And I'm pretty confused if Linux and UNIX are two different entities cuz my college professor seems to use these terms interchangeably and creating all this massive chaos between these two lol. So I did some research at my own to see the difference but the results were quite confusing at first glance, so now I want yall to give me correct answer, are they both same or different????

I had a requirement to implement client server model using ipc. With prerequisite that it should be single threaded and server should handle multiple clients I used Unix domain socket and epoll to achieve this. Now I have another requirement to implement the same using shared memory. How to synchronize between multiple connections about read/write notification? I know I can use semaphore, but I can't use threads for each connection. How to achieve this in single thread process? Using a pipe I can pass the event, but here the performance gain is lost because of the pipe. Any other way I can synchronize?

In approx. 4 hours, we're bumping to 19 decimal years, or 19*1000*86400 Unix seconds.

Here's a few things to remember in 18th dYear: https://0oo.li/event/11001/ .

function uday {
  local unixsec=$(date +%s)
  local unixday=$((unixsec/86400.0))
  printf "%.5f\n" $unixday
}

I’ve worked with C++ and Python and even did web development for a while. Im trying to transition into systems Programming. I’ve taken operating systems and computer architecture courses and they all used c for obvious reasons and I loved those courses.

I’ve been trying to decide which book and loved Stephenson TCP IP Illustrated but I’m also a sucker for no starch press so I’m stuck between which book to choose.

I dislike how LPI doesn’t have exercises, but it’s not a deal beaker

I get the impression from the Unix community that if there is a single other program as a part of software - making it a suite - then it's automatically garbage.

Like, why? Let's take for example systemd. It's an init system, a boot manager, a service manager and DNS resolver (those are the ones I know). What's wrong with it? I believe it actually makes it easier to set them up, since you know they'll integrate with each other pretty well.

Another example: MS Visual Studio (For this example, we're assuming it's available on Linux, but it currently is not, thanks Microsoft)

Visual Studio, a code editor, a debugger, a GUI for git and a compiler. Again, I don't understand what's wrong with it. Sure, it takes up a LOT of space, but nowadays, most people have a 500+ GB drive, but with two downloads and a couple of clicks (3 at most), you'll be up and running with a fully-functioning coding environment, no extra setup needed.

I just feel like the Unix community should be more welcoming when it comes to software suites.

I have a Heathkit ET-3400. If I got one of the redesigned memory expansion boards for it, is it possible to run something like NitrOS-9 on it?

I know a 6800 is no 6809, but I would really like to try and use the ET-3400.

I just want to connect a terminal and use a shell, vi, and assembler if that would somehow be possible.

(Fuzix is a similar project, I know they have disk images for 6809)