Boolean in C

kshegunov

comparing a boolean to true or false is redundant.

Not to mention wrong, generally speaking. 1 evaluates to true but so does -1, so comparing against true is simply the way to break it. Enums implicitly decay to the underlying type so checking against condition and !condition is the correct way to do it, even if a typedef is used.

sierdzio

I find it much easier to understand if something is false when I see if (something == false) than when I see if (!something). Especially in longer expressions it is very easy to miss a single character like ! and read the code wrong.

JonB

@sierdzio
I see a personal-choice-disagreement debate looming... ;-)

I do agree it is "unfortunate" that C chose just that little ! for "not". But personally I never write == false or != false, because I would never "say" that in RL....

J.Hilk

@sierdzio , @JonB

are you guys aware, that not is a valid keyword in c++ ?

https://en.cppreference.com/w/cpp/keyword/not

JonB

@J-Hilk
Yup. And it's devil's-spawn! ;-) [Same for and & or. If I wanted to program in Python or Pascal I would have picked that.] I would never use that, as "nobody" (most people) else uses it or knows about it, so I would regard it as an anti-pattern!

sierdzio

@J-Hilk said in Boolean in C:

@sierdzio , @JonB

are you guys aware, that not is a valid keyword in c++ ?

https://en.cppreference.com/w/cpp/keyword/not

Yes but not in all compilers :-( MSVC does not recognize it.

JonB

@sierdzio
Good, but are you sure? Since it is valid since C99, I would have thought that MSVC would accept those?

sierdzio

@JonB said in Boolean in C:

@sierdzio
Good, but are you sure? Since it is valid since C99, I would have thought that MSVC would accept those?

Last time I tried was last year. Clang, GCC all are 100% fine with it, MSVC was throwing errors.

I now see it's supposed to be defined in some iso646.h header, I never included it, perhaps that's the reason.

J.Hilk

@sierdzio said in Boolean in C:

Yes but not in all compilers :-( MSVC does not recognize it.

VS been nonconforming! 😱 Color me surprised 😉

JonB

@sierdzio
If one has to include a header file for them, makes me wonder if they are not "part of the language", just should be available if you include the header. Are they just #defines in that file??

sierdzio

@JonB said in Boolean in C:

@sierdzio
If one has to include a header file for them, makes me wonder if they are not "part of the language", just should be available if you include the header. Are they just #defines in that file??

yup :D

#define and    &&
#define and_eq &=
#define bitand &
#define bitor  |
#define compl  ~
#define not    !
#define not_eq !=
#define or     ||
#define or_eq  |=
#define xor    ^
#define xor_eq ^=

They are not actual C++ language reserved keywords.

JonB

@sierdzio
Hmmmm.... So does gcc have these in some header file, or does their C++ actually have them as reserved? It does have an iso646.h file, with the #defines, yet you said they worked for you in gcc without you explicitly including that? Does it include it automatically or from something else?

J.Hilk

@JonB said in Boolean in C:

worked for you in gcc without you explicitly including that? Does it include it automatically or from something else

This header was originally in the C standard library as <iso646.h>.
Compatibility header, in C defines alternative operator representations which are keywords in C++.
This means that in a conforming implementation, including this header has no effect.

gcc has mostly a conforming implementation, at least in this regard

JonB

@J-Hilk said in Boolean in C:

This header was originally in the C standard library as <iso646.h>.

So what file is this in, which you say is included automatically?

jsulm

@JonB said in Boolean in C:

which you say is included automatically?

My understanding is that it is NOT included automatically in conforming C++ implementations because those understand these words as keywords and don't need this header file (which is only there for compatibility reasons).

J.Hilk

@JonB I'm talking about the iso646.h @sierdzio mentioned

https://en.cppreference.com/w/cpp/header/ciso646

Probably the reason why MSVC doesn't have those as keywords but requires this header is, IIRC, that it doesn't have/use a dedicated c compiler for c headers and it would break legacy stuff if those were used as keywords

JonB

@jsulm
Mine too. But I quote from @sierdzio above:

Last time I tried was last year. Clang, GCC all are 100% fine with it, MSVC was throwing errors.

I now see it's supposed to be defined in some iso646.h header, I never included it, perhaps that's the reason.

My question is (should be) aimed at him: he says it worked automatically in GCC/Clang (but not MSVC) with no #include from him, that's what I'm trying to understand.

fcarney

I know Boolean is someone's name.
But it kind of look like a diet fad to scare people skinny: boo-lean.

jsulm

@fcarney You mean Anne Boleyn, one of the wifes of king Henry VI which was executed. She was mother of Elisabeth I :-)

jeremy_k

@jsulm said in Boolean in C:

@fcarney You mean Anne Boleyn, one of the wifes of king Henry VI which was executed. She was mother of Elisabeth I :-)

Ha! I'm not sure what a Boleyn variable would be. Dangerous in some manner.

Presumably @fcarney is referring to George Boole.