@cperciva Well said but unfortunately, this race of adding new features is clearly visible in some FOSS projects as well. UNIX philosophy is underrated!
@ngortheone@ed_maste @ariadneconill Actually, the #FreeBSD passwd is just a 100 LoC wrapper around PAM's chauthtok call! OpenBSD implementation is far more sophisticated and secure.
@jcs While there is actually no hard evidence that such changes (Git contribution) are going to change the daily routines of reviewing changes in @OpenBSD_ports, still this argument is the strongest amongst the most nay sayers. It's the fear factor!
I just finished early implementation of macvlan support to my #FreeBSD kernel repo and then asked myself WHY?! Now I think configuring multiple MAC addresses should be done without the need for an external virtual interface! Shouldn't it be done just like aliasing IP addresses?
@jcs Thank you @jcs for recording this Q&A. I really enjoyed watching your answers to the questions. I really hope can do more podcasts in the future.
I could really use some help friends. I'm sorry for the constant RTs. I'm just struggling to get by and am dealing with a lot of constant anxiety and stress in my life, which is not something that's easy to talk about. The pandemic and now recent bad weather aren't helping. 😟
@klarainc At first, it was a "panic" message along with some text and finally the DDB prompt. But that was just the old days. Nowadays I simply fire up my gdb and attach to it. Great opportunity for me to learn about operating system programming, so no worries!
@canadianbryan Sadly I'm not quite familiar with BSD Auth but totally agree that PAM has some serious design flaws like sharing the program's address space with PAM modules. But I can't see why PAM can't be embedded inside a separated privilege child process. That how OpenSSH works, right?
@JackRhysider It is interesting when bad code becomes more tolerable when we pretend they're just-food. And not just any kind of food, but Italian food.