>>107878988
>because they kept fucking themselves over by hardcoding things, over time the linux kernel developers started making more of their syscalls take a variable input (often via a pointer to a struct) alongside its expected size. clever optimization on their end
so clever, good job linuxtards. this is how windows APIs and system calls have worked since the beginning of windows nt. linshitters only started figuring this out decades after they started developing their kernel.

>>107879090
well, you know. better late than never

>>107878988
man7 page for signal(2) has this wonderful line:

typedef typeof(void (int))  *sighandler_t;

which replaces the older line

typedef void (*sighandler_t)(int);

which I think was more readable, desu.
Both are admittedly more readable than the old

void (*signal(int, void (*)(int)))(int);

>>107880368
any time you're doing typedefs of function signatures, readability is going to be a little bit difficult

>>107881327
I think

typedef void (*sighandler_t)(int);

is pretty readable

>>107881427
that's fair. it generally is, i would agree. i do wish C had something like C++ using syntax

>>107880368
Anything that isn't signalfd is a dogshit interface and you shouldn't use it.

>>107878988
when are you going to talk about pledge and unveil? those are the only syscalls i want to learn about

>>107881914
When your retarded ass submits a patch to Linus Torvalds with an implementation of that shit nobody cares about.