With sudo-rs in the news lately, I thought it might be a good time to discuss viable sudo alternatives. I investigated building some alternatives from source last year and ran across some interesting alternatives: Captcha Check doas is very popular with BSD enthusiasts. sup also looks very promising if you like minimalist utilities. I’d be curious to hear if anyone has other sudo alternatives they use regularly or would recommend?
I just su if there’s more than few commands, like when updating. If I only need to mount a drive/similar one command task I use sudo. Never had any issues with sudo so why change to doas.
Exactly the same here.
I’ve taken a look at doas in the past, looks nice and simple (contrary to sudo, with its millions of options), but I didn’t quite see why I would need it. I stuck with sudo.
It might be, but I was never forced to use them. ![]()
I wonder if not using sudo or any of its alternatives is an option.
Can I remove sudo? Does the system need it?
I was always perfectly happy with su on a home computer.
Netinstall of Debian doesn’t include sudo ![]()
As I discovered fairly recently - and had to use “su -” to update and get sudo installed… So on a base netinstall of debian - you need to know the root password…
Maybe, I don’t know. For me it’s more a habit to have it.
For me, it is a habit to use only su. I always put the su window at bottom left and keep it small. I dont walk away and leave the machine with an su running.
Wasn’t nala a wrapper around apt?
I don’t know; I’ve never heard of it. I’m fully OK with apt.
That depends on the options you chose when installing in the first place.
Add sudoers privileges (correct error ‘username’ is not in the sudoers file)
su root
nano /etc/sudoer
then add the user below admin user:
user_name ALL=(ALL) ALL
After which, all will be well and you can use sudo
I guess from the comments so far, no one’s been having issues with the new sudo-rs on their systems.
When are these rust addicts going to leave us alone?
There is no talent involved in translating a program into another language.
When some new algorithms appear, in rust or any language, I might be interested.
Good question, really!
Thing with Rust is that it supposedly eliminates an entire class of security issues. More correctly: you’d need to bend over backwards in order to create one.
The language follows certain principles which are quite helpful. Then there’s things like Option and Result from its STL which are intended to get rid of the issue of NULL.
I think the rewrite of sudo in Rust was triggered by a vulnerability which was found to be around for 10 years already.
The annoying thing about Rust’s STL is that it does do a lot of calls to the panic macro (crash) in failure situations, which is not ok in my book. However, those cases are documented quite well.
Rust is C in a straightjacket.
So why not fix the problem in C?
They need to change the sudo algorithm, not change the language.
No, not OK.
I would like to see an entirely different approach from ths Rust camp. Lets see them develop Rustix from scratch instead of merely translating bits of Linux. Then we could have a fair comparison.
I have no objection to the Rust language… but it needs to achieve something on its own rather than sneak into Linux with mere translations. … Lets see them invent an operating system that takes advantage of Rust’s capabilities.
It already exists
if Rust is so superior, how come Redox has not taken over from Linux?
You could say the same thing about OS/2 - it WAS superior to MS Windows (and DOS + Windows) in almost every way, but didn’t take over from Linux…
It took nearly 10 years for Linux to take over from UNIX (Solaris, AIX, HP-UX etc)…
I’m not standing up for Redox - I barely know anything about it - but marketing/sales success doesn’t mean the product is better than one that didn’t succeed…
Yes,
If there is not much difference, marketing will win.
If one product is really very superior, I would like to think it would win regardless of marketing.
In those early years, Linux really was second best. It was unstable and not secure.
What I really wanted to say was, it there is a bug, the fault is in the algorithm or its implementation, not in the language.
What if a programming language, due to its design, simply prevents a whole class of bugs from happening in the first place. That was one of the dreams Rust was trying to fulfill.
C is an old language, but due to its design it’s extremely easy to make mistakes which result in all kinds of bugs. Now, it’s been the default for a long time (and will probably be for a long time to come, unless Zig replaces it (a language which allows the same time of footnukes as C, but has some advantages over C). I don’t think Rust will replace C, but I do think Rust will eventually surpass C++, due to some really nice mechanisms in Rust, which C++ will simply never have.
I predict the Linux kernel Rust project will fail, because of hostility. However, Rust will take over in other spaces. It just needs to find its niche.
Zig is an interesting language, which is supposed to function as a drop-in replacement for C and C++. It can also interoperate with those two languages without jumping through any hoops. Also, its designers aim for Zig to be a boring language (much like C), where they’re hesitant to add any new features once they’ve reached 1.0. I think this is good for the crowd who wants to have a nice, predictable language which will be the same in 10 years from “now” (whenever that may be). Also, Zig allows the same footnukes as C.