home bbs files messages ]

Forums before death by AOL, social media and spammers... "We can't have nice things"

   comp.os.linux.misc      Linux-specific topics not covered by oth      135,536 messages   

[   << oldest   |   < older   |   list   |   newer >   |   newest >>   ]

   Message 135,225 of 135,536   
   Richard Kettlewell to Chris Ahlstrom   
   Re: Python: A Little Trick For Every Nee   
   04 Feb 26 22:12:39   
   
   From: invalid@invalid.invalid   
      
   Chris Ahlstrom  writes:   
   > Richard Kettlewell wrote this post by blinking in Morse code:   
   >> The Natural Philosopher  writes:   
   >>> All languages are error prone.   
   >>   
   >> They are not all error-prone in _the same way_, and C stands out as   
   >> especially fragile. There are whole classes of vulnerability that either   
   >> don’t exist in other languages or need the programmer to much more   
   >> deliberately go ‘off piste’ before they can happen.   
   >   
   > How about assembler? :-)   
      
   In some ways safer than C. You still need explicit bounds checks (and so   
   on) but if you forget them the outcome is a bit more predictable than in   
   C.   
      
   >>> And blaming that for deficiencies in programmer quality  is just   
   >>> sticking your head in the sand.   
   >>   
   >> I’m not say that there aren’t lazy and incompetent programmers. I   
   >> remember a colleague at a previous job proposing that we could work   
   >> faster by skipping bounds checking in network-facing code, because we   
   >> “knew” what maximum sizes the inputs would be. Obviously in C the   
   >> consequences (had anyone paid attention to that individual) would have   
   >> been vulnerabilites.  In a language with automated bounds checking the   
   >> question wouldn’t even have arisen.   
   >   
   > I dunno, man, the Linux kernel is written and C and it works pretty   
   > well and safely.   
      
   2025 saw over 5,000 CVEs published for the Linux kernel. They do have   
   quite a liberal assignment policy, AIUI any bug relevant to the kernel’s   
   security posture gets a CVE without deeper analysis, but when you have   
   that rate of bugs, you’re not going to attempt a PoC for all of them:   
   the only realistic option is to fix them, log them, and move on to the   
   next one.   
      
   Even before they switched to that policy they’d been doing well over a   
   hundred CVEs per year for some time.   
      
   There’s a reason they’ve put in the effort to enable Rust in the kernel.   
      
   --   
   https://www.greenend.org.uk/rjk/   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)   

[   << oldest   |   < older   |   list   |   newer >   |   newest >>   ]


(c) 1994,  bbs@darkrealms.ca