tags:

views:

321

answers:

4
+5  Q: 

What is !0 in C?

I know that in C, for if statements and comparisons FALSE = 0 and anything else equals true.

Hence,

int j = 40
int k = !j

k == 0 // this is true

My question handles the opposite. What does !0 become? 1? 

int l = 0
int m = !l

m == ? // what is m?
+1  A: 

Generally, yes, it'll become 1. That said even if that is guaranteed behavior (which I'm not sure of) I'd consider code that relied on that to be pretty awful.

You can assume that it's a true value. I wouldn't assume anything more.

Steven Schlansker  2010-09-07 19:01:17
Not just generally; *always*. Boolean operators in C are required to return either 0 or 1.
jamesdlin  2010-09-07 19:05:26
Still, it feels like a violation of having your code show intent.
Steven Schlansker  2010-09-07 19:27:25
Not at all. Things like `prefix="-"+!neg;` are perfectly sane C.
R..  2010-09-07 19:56:40
@R..: I'd not seen that one before. If neg is false, it points to the null byte at the end of the string? Cuter than having prefix be a char which is either 45 or 0, and having the display routine filter out null bytes).
supercat  2010-09-08 14:52:00
Yeah, if `neg` is false then `prefix` points to an empty string. If it's true, `prefix` points to the string `"-"` (with null termination either way, of course).
R..  2010-09-08 16:18:24
+1  A: 

The Bang operator (!) is the logical not operator found commonly in C, C++ and C#, so 

!0 == 1
!1 == 0

This is based on the language characteristic of what is interpreted to be either true or false... in more modern languages it would be like this

!false == true
!true == false

See DeMorgan Law concerning truth tables...

tommieb75  2010-09-07 19:08:02
Actually, in C++, `!` takes a `bool` operand and produces a `bool` result -- though for backward compatibility, there are implicit conversions from `bool` to `int`, with `true` converting to `1` and `false` to `0`, and from other integer types to `bool`, with `0` converting to `false`, and any other value converting to `true`.
Jerry Coffin  2010-09-07 19:14:35
+20  A: 

Boolean/logical operators in C are required to yield either 0 or 1.

From section 6.5.3.3/5 of the ISO C99 standard:

The result of the logical negation operator ! is 0 if the value of its operand compares unequal to 0, 1 if the value of its operand compares equal to 0.

In fact, !!x is a common idiom for forcing a value to be either 0 or 1 (I personally prefer x != 0, though).

Also see Q9.2 from the comp.lang.c FAQ.

jamesdlin  2010-09-07 19:08:57
Good call. Looks like more of us need to actually read the damn ISO Standard.
Justin Niessner  2010-09-07 19:15:22
+1 for pointing to the standard, but code that uses `!!` needs to head straight to the garbage can.
Billy ONeal  2010-09-07 19:17:01
This should be the accepted answer.
Stephen Canon  2010-09-07 19:19:02
Of course not a lot of C compilers feel particularly compelled to follow the ISO standard. GCC probably, but with anyone else its a crapshoot.
T.E.D.  2010-09-07 19:20:33
@Stephen Canon - Well, vote it up enough and he'll at least get the Populist badge. It's a gold too.
T.E.D.  2010-09-07 19:21:21
Please link to the ISO standard, so that we too can look it up sometimes.
BiGYaN  2010-09-07 19:21:37
GCC, ICC, clang, the IBM and HP compiler tools all make a good faith effort to follow the standard. I'm sure I'm forgetting some others. MSVC is the only compiler I've encountered that doesn't at least *try* to follow the standard.
Stephen Canon  2010-09-07 19:22:35
@T.E.D.: Sadly, the accepted answer needs to have at least 11 votes for him to get populist.
Stephen Canon  2010-09-07 19:23:22
@BiGYaN: AFAIK the C standard must be purchased, and I am not aware of a legitimate, free version that I could link to. Also see http://c-faq.com/ansi/avail.html
jamesdlin  2010-09-07 19:25:52
Curious: Is this true in C89?
mathepic  2010-09-07 19:29:51
@mathepic: Yes, it's true for C89 as well (section 6.3.3.3).
jamesdlin  2010-09-07 19:38:30
@Billy ONeal: To my mind, saying "flagvar = !!(multibitFlag " is clearer than "flagvar = ((multibitFlag " What is gained by the extra verbosity of the latter format?
supercat  2010-09-07 20:02:44
RBerteig  2010-09-07 22:23:45
@supercat: What is gained is the clarity of the code. `!!` is trying to be tricky, `!= 0` explains explicitly to the reader what the code is doing. However, in that case, there's no reason for the `!!` or `!=` at all.
Billy ONeal  2010-09-08 00:20:27
@Billy ONeal: `!!` is perfectly clear to an experienced C programmer, since it's the idiomatic way of flattening a value to 1 or 0. `!= 0` is no better since it relies on knowing that the result of the `!=` operator is either 1 or 0, as opposed to the `!` operator. In fact I'd argue that it's worse, since `!!` more clearly conveys the intent. If you want a better alternative, I suggest `? 1 : 0`.
caf  2010-09-08 02:22:10
Or use a C99 toolchain, `#include <stdbool.h>` and cast to `bool`.
Stephen Canon  2010-09-08 03:09:07
@Stephen Canon: Is a cast of any numeric type to "bool" guaranteed to convert any non-zero value to 1? Some compilers I use have a "bit" type which behaves as a one-bit integer (so assignments just use the LSB of the source operand).
supercat  2010-09-08 14:42:46
@Billy ONeal: I find the "!!" format more readable, since there's less stuff to visually parse, and to my mind the concept of "nonzero" is distinct from the concept of "doesn't equal the numerical constant zero". And there is a reason for the "!!": if multiBitFlag grows to be bigger than flagVar, the "!!" ensures the result will fit. It's also often nicer for routines' contracts to indicate that they'll return 1 or 0, rather than just some arbitrary non-zero value or zero (though I know that e.g. strcmp etc. merely specify the sign of their results, and not the value).
supercat  2010-09-08 14:48:13
@supercat: "When any scalar value is converted to _Bool, the result is 0 if the value compares equal to 0; otherwise, the result is 1." (§6.3.1.2)
Stephen Canon  2010-09-08 15:13:21
jamesdlin  2010-09-08 17:12:47
+3  A: 

§6.5.3.3/5: "The result of the logical negation operator ! is 0 if the value of its operand compares unequal to 0, 1 if the value of its operand compares equal to 0. The result has type int."

The other logical operators (e.g., &&, ||) always produce either 0 or 1 as well.

Jerry Coffin  2010-09-07 19:10:31

related questions

any good tool for makefile generation?
How do you pass a function as a parameter in C?
Where should a veteran C programmer start in order to master Java ?
Any good book on best practice and guidelines in developing a SDK in C?
How do you determine the size of a file in C?
Decoding printf statements in C (Printf Primer)
Shift operator in C
How to avoid redefining VERSION, PACKAGE, etc.
Alpha blending sprites in Nintendo DS Homebrew
When should I use type abstraction in embedded systems
How to implement continuations?
What are the barriers to understanding pointers and what can be done to overcome them?
Anyone have experience creating a shared library in MATLAB?
String.indexOf function in C
Passing multidimensional arrays as function arguments in C
C/C++ library for reading MIDI signals from a USB MIDI device
Choosing a static code analysis tool
How do you printf an unsigned long long int?
Good STL-like library for C.
Rockbox audio format
Why am I getting a malloc: double free error with realloc()?
Should I learn C?
GTK implementation of MessageBox
Is gettimeofday() guaranteed to be of microsecond resolution?
How to use the C socket API in C++ on z/OS