Given the “conversation” around rust this title is effectively a lie by omission. He is stepping down due to a lack of time not anything wrong with rust or kernel development. Given the kernels state of maintainer drought from too many devs this is barely news
this post was submitted on 24 Nov 2025
115 points (96.7% liked)
Linux
10265 readers
784 users here now
A community for everything relating to the GNU/Linux operating system (except the memes!)
Also, check out:
Original icon base courtesy of lewing@isc.tamu.edu and The GIMP
founded 2 years ago
MODERATORS
In this thread: people who believe the myth of the safe C programmer. The one who has memorized the spec and is able to hold endless context in their brain while writing code. They themselves are C compilers.
You're kidding yourself if you think most people in this thread are actually programmers. Most people here wouldn't know the difference between C and Rust code if they saw it, let alone be able to write anything in either language.
I think that there are certain attitudes that mainly occur to people outside the domain. Like how people endlessly shit on open-source projects, but few of those people are ever actually at the wheel of one.
the myth of the safe C programmer
I learnt C around about 1997 and I've used it off and on professionally since about 2006. I am not a myth, and there are many others like me.
What do you want me to write?
What do you want me to write?
To meet the bar set by onlinepersona, you'd need to write safe C code, not just some of the time, but all of the time. What you appear to be proposing is to provide evidence that you can write safe C code some of the time.
It's like if somebody said "everyone gets sick!", and some other person stepped up and said "I never get sick. As proof, you can take my temperature right now; see, I'm healthy!". Obviously, the evidence being offered is insufficient to refute the claim being made by the first person
I want you to write kernel code for a few years. But we go to Lemmy with the machismo we have, not the machismo we wish we had. Write a JSON recognizer; it should have the following signature and correctly recognize ECMA 404, returning 0 on success and 1 on failure.
int recognizeJSON(const char*);
I estimate that this should take you about 120 lines of code. My prior estimated defect rate for C programs is about one per 60 lines. So, to get under par, your code should have fewer than two bugs.
Can you point to relevant non-trivial public work of yours that has zero CVE's?
The more you learn and know, the more you refrain from making such statements. This is universally applicable, and not limited to C or programming. And that's what makes your "story" suspect.
Or maybe it's a reading comprehension issue.
^I^ ^used^ ^to^ ^write^ ^non-trivial^ ^C^ ^code^ ^myself^ ^btw.^
That's quite an insulting insinuation, and no, I'm not going to doxx myself on my pseudonymous piefed account.
What do you want me to write?
Super-human claims require evidence. And asking for that evidence is not an insult.
I think it's time for this instance to consider introducing a filter where users have to choose a language they know (any language), and then have to answer easy questions about it (in a specific way), before being able to post here.
It can be limited to specific posts, to limit the false-negative filtering of genuine discourse.
This should help with bots, or worse, actual humans who accepted being shaped into acting like ones. The line separating the two has become very thin anyway, given the prevalence of LLM use, both automatic AND manual.
