« first day (667 days earlier)   

10:12 AM
@Dennis Can you pull Add++ please?
 
 
4 hours later…
2:11 PM
@ngn What happened to the ngn/k repo? I can no longer access it.
@cairdcoinheringaahing Done.
 
@Dennis Thanks!
@Dennis I found another bug, would you mind pulling it again please?
 
Done.
 
ngn
@Dennis ah, I didn't know you'd be updating it now. access restored
 
2:29 PM
@ngn I wasn't updating, I got an email saying I no longer had access.
That wasn't very clear. I shouldn't write chat messages before I wake up...
It's fine if I can't access it temporarily. I was just wondering what was going on.
 
ngn
@Dennis i'm being paranoid for no good reason :)
 
@Dennis that looks like good advice to me, yeah
 
@ngn Ah, OK. I don't really need access to the source code. A pre-built ELF would be plenty.
 
ngn
2:45 PM
@Dennis that sounds like a good idea
though, to be honest, figuring out what my code does would take as much effort as studying the disassember's output :)
 
Probably. :P
 
ngn
@Dennis but you shouldn't trust my executable (though I don't intend to do anything naughty) - how hard would it be to run it with some limitations?
for instance chroot to a particular folder, no network access, virtual memory limit
 
The executable runs only in the sandbox.
People can already execute raw ELF files via tio.run/#elf, in a severely restricted SELinux context rather than chroot. Adding a closed-source ELF is not a security risk.
 
ngn
@Dennis they can download stuff: tio.run/##y9bNS8/7/9/QSkk/v6BEPxvE1c9W@v8fAA
 
Only files the sandbox context has access to.
That has to include the interpreters, or the sandbox wouldn't be able to execute them.
 
ngn
2:59 PM
@Dennis right, I was just thinking about that... what deamons usually do is create their own chroot after the process has started, but that would be too hard for me to implement, and also pointless
 
Most ways of communicating with a daemon aren't available in the sandbox anyway.
 

« first day (667 days earlier)