C99 or C11?

Serious topic. No mention of C89 as it is objectively inferior to C99.

>>10,11
I mostly agree with Linus on bools (see https://lkml.org/lkml/2013/8/31/138 and ), in that they aren't strictly defined enough for a programmer to use them better than `This is a char or maybe an int depending on how my compiler feels. It is probably zero or nonzero,' and you don't really get any benefit (in packing in structs, for example) that a compiler can't also give to a sanely-used integer type.

In addition, I've never really seen a situation in which I wanted to use bool in which there was a downside to using char or even int instead, because in the place I wanted to use the variable, it was always possible for me to tack on more information like an error code, or some useful metric that divided all inputs into two partitions, etc. That's just my experience, and possibly there exists an example that would make me change my mind, but I doubt it.

I would normally give weight to your clarity argument, but the following compiles for me just fine ([m]gcc -o a a.c -std=c99 -Wall -Werror[/m]):

#include <stdio.h>
#include <stdbool.h>

int main(void)
{
bool t, u;
t = 2;
u = t - 2;
if (u) printf("meh\n");

return 0;
}

and outputs

meh

I know perfectly well why it does so, and I won't say that this program is somehow wrong, but I believe it underscores that in a language like C, bool doesn't really help the programmer avoid performing numerical operations on the values.

But really, the reason I picked on stdbool.h specifically is that a lot of C programmers (like Linus) point to it as proof of C++'s insidious influence upon C.