I am aware of two techniques for flipping a bool:
x = !x
and
x ^= 1
I prefer the second, in the same way that I prefer x++ over x+=1 over x=x+1.
But are they guaranteed to produce the same Assembly code?
And if not, is there some rationale to favour one over the other?
There are never any guarantees about same assembly code. Yes, given bool x, they both do exactly the same thing. Yes, that implies they are likely to be the same.
It is a common fallacy that writing expressions in an unusual way might make them faster. Avoid working with anyone who makes habits based on such ideas.
x = ! x is clearer because ! is defined in terms of Boolean values.
I prefer the second, in the same way that I prefer x++ over x+=1 over x=x+1.
Please prefer the first. It is expected and easily understood by most programmers. The second form, only works for values 0 and 1 in integer's case.
But are they guaranteed to produce the same Assembly code?
No.
And if not, is there some rationale to favour one over the other?
Decreasing the ratio of "wtf/loc" of you code (i.e. how many times would another developer look at your code and say "WTF?!?", over every n lines of code).
Edit: Either way, you will probably never see a real-world example of an application made faster by replacing one expression with the other. It is a matter of style, not a matter of performance.
The second expression is simply a source of difficulty found bugs. For example you may think that some expression has type bool while its actual type due to the integer promotion or usual arithmetic conversion is some integral type.
Consider the following code snippet
unsigned int x = 3;
x ^= 1;
std::cout << x << std::endl;
if ( x == false ) std::cout << "All is O'k\n";
else std::cout << "Oops. Something is wrong!\n";
x = 3;
x = !x;
std::cout << x << std::endl;
if ( x == false ) std::cout << "All is O'k\n";
else std::cout << "Oops. Something is wrong!\n";
So using expression x ^= 1; to flip a boolean value can only confuse readers of the code.
So I would prefer to use x = !x; instead of x ^= 1;
Do not forget about the principle KISS: Keep It Simple Stupid.:)
Though for example in C the result of operator ! has type int nevertheless the value of the operation is either 0 or 1.
From the C Standard
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.
So for both languages there is a guarantee that the result will be exactly either 0 or 1 for any source value.
!x and x^1 are clearly different operations.
The former is a logical negation, returning a 0/1 value. It makes sense to use it on bools.
The latter is a bitwise exclusive or, returning any possible value. It shouldn't be used on a bool. (For xoring bools, prefer the inequality comparison !=). It is also less efficient in unoptimized code as it invoves and extra operand (1).
K&R forgot to provide a !! operator, that you would have loved, and possibly ~~.
Related
I was messing around with some c++ code after learning that you cannot increment booleans in the standard. I thought incrementing booleans would be useful because I like compacting functions if I can and
unsigned char b = 0; //type doesn't really matter assuming no overflow
while (...) {
if (b++) //do something
...}
is sometimes useful to have, but it would be nice to not have to worry about integer overflow. Since booleans cannot take on any value besides 0 or 1, I thought that perhaps this would work. Of course, you could just have a boolean assigned to 1 after you do your operation but that takes an extra line of code.
Anyway, I thought about how I might achieve this a different way and I came up with this.
bool b = 0;
while (...)
if (b & (b=1)) ...
This turned out to always evaluate to true, even on the first pass through. I thought - sure, its just doing the bitwise operator from right to left, except that when I swapped the order, it did the same thing.
So my question is, how is this expression being evaluated so that it simplifies to always being true?
As an aside, the way to do what I wanted is like this I guess:
bool b = 0;
while (...) if (!b && !(b=1)) //code executes every time except the first iteration
There's a way to get it to only execute once, also.
Or just do the actually readable and obvious solution.
bool b = 0;
while (...) {
if (b) {} else {b=1;}
}
This is what std::exchange is for:
if (std::exchange(b, true)) {
// b was already true
}
As mentioned, sequencing rules mean that b & (b=1) is undefined behaviour in both C and C++. You can't modify a variable and read from it "without proper separation" (e.g., as separate statements).
There is no "strange behavior" here. The & operator, in your case is a bitwise arithmetic operator. These operator are commutative, so both operand must be evaluated before computing the resulting value.
As mentioned earlier, this is considered as undefined behavior, as the standard does not prevent computing + loading into a register the left operand, before doing the same to the second, or doing so in reverse order.
If you are using C++14 or above, you could use std::exchange do achieve this, or re-implement it quite easily using the following :
bool exchange_bool(bool &var, bool &&val) {
bool ret = var;
var = val;
return ret;
}
If you end up re-implementing it, consider using template instead of hard-coded types.
As a proper answer to your question :
So my question is, how is this expression being evaluated so that it simplifies to always being true?
It does not simplify, if we strictly follow the standard, but I'd guess you are always using the same compiler, which end up producing the same output.
Side point: You should not "compact" functions. Most of the time the price of this is less readability. This might be fine if you're the only dev working on a project, but on bigger project, or if you come back to your code later, that could be an issue. Last but not least, "compact" function are more error prone, because of their lack of readability.
If you really want shorter function, you should consider using sub-function that will handle part of the function computation. This would enable you to write short function, to quickly fix or update behavior, without impacting clarity.
I don't see what meaning you can assign to a "bitwise increment" operator. If I am right, what you achieve is just setting the low-order bit, which is done by b|= 1 (following the C conventions, this could have been defined as a unary b||, but the || was assigned another purpose). But I don't feel that this is a useful bitwise operation, and it can increment only once. IMO the standard increment operator is more bitwise.
If your goal was in fact to "increment" a boolean variable, i.e. set it to true (prefix or postfix), b= b || true is a way. (There is no "logical OR assignement" b||= true, and even less a "logical increment" b||||.)
NaN propagates through "most" operations as described in NaN - Wikipedia.
I would like to know the operations which NaN does NOT propagate. For example, I'm coding in C++ and found that following code prints 1, which is not NaN.
const double result = std::pow(1, std::numeric_limits<double>::quiet_NaN());
std::cout << result << std::endl;
For std::powfunction, this behavior is described in std::pow - cppreference.com.
Could you share any other examples?
Here's an example demonstrating functions of NaN which return non-NaN. The list is in IEEE 754-2008, 9.2.1 Special values (there are some others functions, but they don't seem to be implemented in C++):
#include <cmath>
#include <limits>
#include <iostream>
int main()
{
const auto nan=std::numeric_limits<double>::quiet_NaN();
const auto inf=std::numeric_limits<double>::infinity();
std::cout << std::hypot(nan,inf) << '\n';
std::cout << std::hypot(inf,nan) << '\n';
std::cout << std::pow(nan, 0) << '\n';
std::cout << std::pow(1,nan) << '\n';
}
The output is:
inf
inf
1
1
std::pow is not really an operator in the sense that something like the multiplication in a * b is an operation. It's a function. And functions can have branches which can deal with a NaN as they please. Another similar std function is, in a sense, std::is_nan. This doesn't propagate NaN either: it returns a bool which is implicitly convertible to a numeric type:
std::is_nan(std::is_nan(a))
is false for any type a that allows the expression to compile. Other examples include std::fpclassify and std::isfinite. I particularly like std::fpclassify as an example since it has an int return type, so really it's no less of a numeric function than pow!
Note that the implicit conversion of NaN to bool is defined to be false true according to the accepted answer to this question, which is important. It allows !, &&, and || to work with an NaN. Finally, the expression separator operator , doesn't propagate NaN either, and an NaN on the unused branch of a ternary conditional operation is ignored.
The other case I would try is pow(0, NaN). If one has pow(0,0) == 1 then one should expect
pow(0, NaN) == 1
There is a rationale for this, and actually it is needed for consistent behavior.
Even though there is no IEEE standard dictating the behavior of all elementary functions with respect to NaN, there is a very basic rule:
If for all finite point numbers x including +inf and -inf (but excluding NaN) we have
f(const1, x) == const2
then (and only then) one also must return a non-NaN result
f(const1, NaN) == const2
#Long explanation
This is because NaN represents "undefined" or "any other numeric value from -inf .. inf". Why? Consider the prototypical example 0/0.
If one has the equation
b = a * x
and wants to solve for x. Clearly the solution is
x = b/a
Now if a == b == 0 then the original equation has infinitely many solutions, namely all finite numbers x. That's why NaN means "unspecified" and 0/0 == NaN. Now if an unspecified number (i.e. NaN) enters a function as an argument, it most often will cause an "unspecified" answer. The exception is where the output does not depend on the input, and in this case one should/must not return NaN.
Consider
pow(1, a/b)
This expression always evaluates to 1 for nonzero a and b, which makes sense from a mathematical point of view as the 1^x in a mathematical sense does not depend on x. So also numerically one would require that for a = 0 or b = 0 (and hence x=NaN) one also obtains 1.
So if one agrees upon the fact that
pow(1,-inf) = pow(1,inf) = 1
then one also has to define
pow(1,NaN) = 1
for consistent behavior. The same applies to pow(0,c) but also evaluation of a polynomial of degree zero should generate a non-NaN output on NaN input.
Note: this reasoning can be applied to any function, including the built-in operators.
I was writing a console application that would try to "guess" a number by trial and error, it worked fine and all but it left me wondering about a certain part that I wrote absentmindedly,
The code is:
#include <stdio.h>
#include <stdlib.h>
int main()
{
int x,i,a,cc;
for(;;){
scanf("%d",&x);
a=50;
i=100/a;
for(cc=0;;cc++)
{
if(x<a)
{
printf("%d was too big\n",a);
a=a-((100/(i<<=1))?:1);
}
else if (x>a)
{
printf("%d was too small\n",a);
a=a+((100/(i<<=1))?:1);
}
else
{
printf("%d was the right number\n-----------------%d---------------------\n",a,cc);
break;
}
}
}
return 0;
}
More specifically the part that confused me is
a=a+((100/(i<<=1))?:1);
//Code, code
a=a-((100/(i<<=1))?:1);
I used ((100/(i<<=1))?:1) to make sure that if 100/(i<<=1) returned 0 (or false) the whole expression would evaluate to 1 ((100/(i<<=1))?:***1***), and I left the part of the conditional that would work if it was true empty ((100/(i<<=1))? _this space_ :1), it seems to work correctly but is there any risk in leaving that part of the conditional empty?
This is a GNU C extension (see ?: wikipedia entry), so for portability you should explicitly state the second operand.
In the 'true' case, it is returning the result of the conditional.
The following statements are almost equivalent:
a = x ?: y;
a = x ? x : y;
The only difference is in the first statement, x is always evaluated once, whereas in the second, x will be evaluated twice if it is true. So the only difference is when evaluating x has side effects.
Either way, I'd consider this a subtle use of the syntax... and if you have any empathy for those maintaining your code, you should explicitly state the operand. :)
On the other hand, it's a nice little trick for a common use case.
This is a GCC extension to the C language. When nothing appears between ?:, then the value of the comparison is used in the true case.
The middle operand in a conditional expression may be omitted. Then if the first operand is nonzero, its value is the value of the conditional expression.
Therefore, the expression
x ? : y
has the value of x if that is nonzero; otherwise, the value of y.
This example is perfectly equivalent to
x ? x : y
In this simple case, the ability to omit the middle operand is not especially useful. When it becomes useful is when the first operand does, or may (if it is a macro argument), contain a side effect. Then repeating the operand in the middle would perform the side effect twice. Omitting the middle operand uses the value already computed without the undesirable effects of recomputing it.
& has &&. | has ||. Why doesn't ^ have ^^?
I understand that it wouldn't be short-circuiting, but it would have different semantics. In C, true is really any non-zero value. Bitwise XOR is not always the same thing as logical XOR:
int a=strcmp(str1,str2);// evaluates to 1, which is "true"
int b=strcmp(str1,str3);// evaluates to 2, which is also "true"
int c=a ^^ b; // this would be false, since true ^ true = false
int d=a ^ b; //oops, this is true again, it is 3 (^ is bitwise)
Since you can't always rely on a true value being 1 or -1, wouldn't a ^^ operator be very helpful? I often have to do strange things like this:
if(!!a ^ !!b) // looks strange
Dennis Ritchie answers
There are both historical and practical reasons why there is no ^^ operator.
The practical is: there's not much use for the operator. The main point of && and || is to take advantage of their short-circuit evaluation not only for efficiency reasons, but more often for expressiveness and correctness.
[...]
By contrast, an ^^ operator would always force evaluation of both arms of the expression, so there's no efficiency gain. Furthermore, situations in which ^^ is really called for are pretty rare, though examples can be created. These situations get rarer and stranger as you stack up the operator--
if (cond1() ^^ cond2() ^^ cond3() ^^ ...) ...
does the consequent exactly when an odd number of the condx()s are true. By contrast, the && and || analogs remain fairly plausible and useful.
Technically, one already exists:
a != b
since this will evaluate to true if the truth value of the operands differ.
Edit:
Volte's comment:
(!a) != (!b)
is correct because my answer above does not work for int types. I will delete mine if he adds his answer.
Edit again:
Maybe I'm forgetting something from C++, but the more I think about this, the more I wonder why you would ever write if (1 ^ 2) in the first place. The purpose for ^ is to exclusive-or two numbers together (which evaluates to another number), not convert them to boolean values and compare their truth values.
This seems like it would be an odd assumption for a language designer to make.
For non-bool operands, I guess what you would want is for a ^^ b to be evaluated as:
(a != 0) ^ (b != 0)
Well, you have the above option and you have a few options listed in other answers.
The operator ^^ would be redundant for bool operands. Talking only about boolean operands, for the sake of argument, let's pretend that ^ was bitwise-only and that ^^ existed as a logical XOR. You then have these choices:
& - Bitwise AND -- always evaluates both operands
&& - Logical AND -- does not always evaluate both operands
| - Bitwise OR -- always evaluates both operands
|| - Logical OR -- does not always evaluate both operands
^ - Bitwise XOR -- must always evaluate both operands
^^ - Logical XOR -- must always evaluate both operands
Why didn't they create ^^ to essentially convert numerical values into bools and then act as ^? That's a good question. Perhaps because it's more potentially confusing than && and ||, perhaps because you can easily construct the equivalent of ^^ with other operators.
I can't say what was in the heads of Kernighan and Ritchie when they invented C, but you made a brief reference to "wouldn't be short-circuiting", and I'm guessing that's the reason: It's not possible to implement it consistently. You can't short-circuit XOR like you can AND and OR, so ^^ could not fully parallel && and ||. So the authors might well have decided that making an operation that sort of kind of looks like its parallel to the others but isn't quite would be worse than not having it at all.
Personally, the main reason I use && and || is for the short-circuit rather than the non-bitwise. Actually I very rarely use the bitwise operators at all.
Another workaround to the ones posted above (even if it requires another branch in the code) would be:
if ( (a? !b : b ) )
that is equivalent to xor.
In Java the ^ operator indeed does do logical XOR when used on two boolean operands (just like & and | in Java do non-short-circuiting logical AND and OR, respectively, when applied to booleans). The main difference with C / C++ is that C / C++ allows you to mix integers and booleans, whereas Java doesn't.
But I think it's bad practice to use integers as booleans anyway. If you want to do logical operations, you should stick to either bool values, or integers that are either 0 or 1. Then ^ works fine as logical XOR.
An analogous question would be to ask, how would you do non-short-circuiting logical AND and OR in C / C++? The usual answer is to use the & and | operators respectively. But again, this depends on the values being bool or either 0 or 1. If you allow any integer values, then this does not work either.
Regardless of the case for or against ^^ as an operator, you example with strcmp() sucks. It does not return a truth value (true or false), it returns a relation between its inputs, encoded as an integer.
Sure, any integer can be interpreted as a truth value in C, in which case 0 is "false" and all other values are "true", but that is the opposite of what strcmp() returns.
Your example should begin:
int a = strcmp(str1, str2) == 0; // evaluates to 0, which is "false"
int b = strcmp(str1, str3) == 0; // evaluates to 0, which is also "false"
You must compare the return value with 0 to convert it to a proper boolean value indicating if the strings were equal or not.
With "proper" booleans, represented canonically as 0 or 1, the bitwise ^ operator works a lot better, too ...
Is there any reason not to use the bitwise operators &, |, and ^ for "bool" values in C++?
I sometimes run into situations where I want exactly one of two conditions to be true (XOR), so I just throw the ^ operator into a conditional expression. I also sometimes want all parts of a condition to be evaluated whether the result is true or not (rather than short-circuiting), so I use & and |. I also need to accumulate Boolean values sometimes, and &= and |= can be quite useful.
I've gotten a few raised eyebrows when doing this, but the code is still meaningful and cleaner than it would be otherwise. Is there any reason NOT to use these for bools? Are there any modern compilers that give bad results for this?
|| and && are boolean operators and the built-in ones are guaranteed to return either true or false. Nothing else.
|, & and ^ are bitwise operators. When the domain of numbers you operate on is just 1 and 0, then they are exactly the same, but in cases where your booleans are not strictly 1 and 0 – as is the case with the C language – you may end up with some behavior you didn't want. For instance:
BOOL two = 2;
BOOL one = 1;
BOOL and = two & one; //and = 0
BOOL cand = two && one; //cand = 1
In C++, however, the bool type is guaranteed to be only either a true or a false (which convert implicitly to respectively 1 and 0), so it's less of a worry from this stance, but the fact that people aren't used to seeing such things in code makes a good argument for not doing it. Just say b = b && x and be done with it.
Two main reasons. In short, consider carefully; there could be a good reason for it, but if there is be VERY explicit in your comments because it can be brittle and, as you say yourself, people aren't generally used to seeing code like this.
Bitwise xor != Logical xor (except for 0 and 1)
Firstly, if you are operating on values other than false and true (or 0 and 1, as integers), the ^ operator can introduce behavior not equivalent to a logical xor. For example:
int one = 1;
int two = 2;
// bitwise xor
if (one ^ two)
{
// executes because expression = 3 and any non-zero integer evaluates to true
}
// logical xor; more correctly would be coded as
// if (bool(one) != bool(two))
// but spelled out to be explicit in the context of the problem
if ((one && !two) || (!one && two))
{
// does not execute b/c expression = ((true && false) || (false && true))
// which evaluates to false
}
Credit to user #Patrick for expressing this first.
Order of operations
Second, |, &, and ^, as bitwise operators, do not short-circuit. In addition, multiple bitwise operators chained together in a single statement -- even with explicit parentheses -- can be reordered by optimizing compilers, because all 3 operations are normally commutative. This is important if the order of the operations matters.
In other words
bool result = true;
result = result && a() && b();
// will not call a() if result false, will not call b() if result or a() false
will not always give the same result (or end state) as
bool result = true;
result &= (a() & b());
// a() and b() both will be called, but not necessarily in that order in an
// optimizing compiler
This is especially important because you may not control methods a() and b(), or somebody else may come along and change them later not understanding the dependency, and cause a nasty (and often release-build only) bug.
The raised eyebrows should tell you enough to stop doing it. You don't write the code for the compiler, you write it for your fellow programmers first and then for the compiler. Even if the compilers work, surprising other people is not what you want - bitwise operators are for bit operations not for bools.
I suppose you also eat apples with a fork? It works but it surprises people so it's better not to do it.
I think
a != b
is what you want
Disadvantages of the bitlevel operators.
You ask:
“Is there any reason not to use the bitwise operators &, |, and ^ for "bool" values in C++? ”
Yes, the logical operators, that is the built-in high level boolean operators !, && and ||, offer the following advantages:
Guaranteed conversion of arguments to bool, i.e. to 0 and 1 ordinal value.
Guaranteed short circuit evaluation where expression evaluation stops as soon as the final result is known.
This can be interpreted as a tree-value logic, with True, False and Indeterminate.
Readable textual equivalents not, and and or, even if I don't use them myself.
As reader Antimony notes in a comment also the bitlevel operators have alternative tokens, namely bitand, bitor, xor and compl, but in my opinion these are less readable than and, or and not.
Simply put, each such advantage of the high level operators is a disadvantage of the bitlevel operators.
In particular, since the bitwise operators lack argument conversion to 0/1 you get e.g. 1 & 2 → 0, while 1 && 2 → true. Also ^, bitwise exclusive or, can misbehave in this way. Regarded as boolean values 1 and 2 are the same, namely true, but regarded as bitpatterns they're different.
How to express logical either/or in C++.
You then provide a bit of background for the question,
“I sometimes run into situations where I want exactly one of two conditions to be true (XOR), so I just throw the ^ operator into a conditional expression.”
Well, the bitwise operators have higher precedence than the logical operators. This means in particular that in a mixed expression such as
a && b ^ c
you get the perhaps unexpected result a && (b ^ c).
Instead write just
(a && b) != c
expressing more concisely what you mean.
For the multiple argument either/or there is no C++ operator that does the job. For example, if you write a ^ b ^ c than that is not an expression that says “either a, b or c is true“. Instead it says, “An odd number of a, b and c are true“, which might be 1 of them or all 3…
To express the general either/or when a, b and c are of type bool, just write
(a + b + c) == 1
or, with non-bool arguments, convert them to bool:
(!!a + !!b + !!c) == 1
Using &= to accumulate boolean results.
You further elaborate,
“I also need to accumulate Boolean values sometimes, and &= and |=? can be quite useful.”
Well, this corresponds to checking whether respectively all or any condition is satisfied, and de Morgan’s law tells you how to go from one to the other. I.e. you only need one of them. You could in principle use *= as a &&=-operator (for as good old George Boole discovered, logical AND can very easily be expressed as multiplication), but I think that that would perplex and perhaps mislead maintainers of the code.
Consider also:
struct Bool
{
bool value;
void operator&=( bool const v ) { value = value && v; }
operator bool() const { return value; }
};
#include <iostream>
int main()
{
using namespace std;
Bool a = {true};
a &= true || false;
a &= 1234;
cout << boolalpha << a << endl;
bool b = {true};
b &= true || false;
b &= 1234;
cout << boolalpha << b << endl;
}
Output with Visual C++ 11.0 and g++ 4.7.1:
true
false
The reason for the difference in results is that the bitlevel &= does not provide a conversion to bool of its right hand side argument.
So, which of these results do you desire for your use of &=?
If the former, true, then better define an operator (e.g. as above) or named function, or use an explicit conversion of the right hand side expression, or write the update in full.
Contrary to Patrick's answer, C++ has no ^^ operator for performing a short-circuiting exclusive or. If you think about it for a second, having a ^^ operator wouldn't make sense anyway: with exclusive or, the result always depends on both operands. However, Patrick's warning about non-bool "Boolean" types holds equally well when comparing 1 & 2 to 1 && 2. One classic example of this is the Windows GetMessage() function, which returns a tri-state BOOL: nonzero, 0, or -1.
Using & instead of && and | instead of || is not an uncommon typo, so if you are deliberately doing it, it deserves a comment saying why.
Patrick made good points, and I'm not going to repeat them. However might I suggest reducing 'if' statements to readable english wherever possible by using well-named boolean vars.For example, and this is using boolean operators but you could equally use bitwise and name the bools appropriately:
bool onlyAIsTrue = (a && !b); // you could use bitwise XOR here
bool onlyBIsTrue = (b && !a); // and not need this second line
if (onlyAIsTrue || onlyBIsTrue)
{
.. stuff ..
}
You might think that using a boolean seems unnecessary, but it helps with two main things:
Your code is easier to understand because the intermediate boolean for the 'if' condition makes the intention of the condition more explicit.
If you are using non-standard or unexpected code, such as bitwise operators on boolean values, people can much more easily see why you've done this.
EDIT: You didnt explicitly say you wanted the conditionals for 'if' statements (although this seems most likely), that was my assumption. But my suggestion of an intermediate boolean value still stands.
Using bitwise operations for bool helps save unnecessary branch prediction logic by the processor, resulting from a 'cmp' instruction brought in by logical operations.
Replacing the logical with bitwise operations (where all operands are bool) generates more efficient code offering the same result. The efficiency ideally should outweigh all the short-circuit benefits that can be leveraged in the ordering using logical operations.
This can make code a bit un-readable albeit the programmer should comment it with reasons why it was done so.
IIRC, many C++ compilers will warn when attempting to cast the result of a bitwise operation as a bool. You would have to use a type cast to make the compiler happy.
Using a bitwise operation in an if expression would serve the same criticism, though perhaps not by the compiler. Any non-zero value is considered true, so something like "if (7 & 3)" will be true. This behavior may be acceptable in Perl, but C/C++ are very explicit languages. I think the Spock eyebrow is due diligence. :) I would append "== 0" or "!= 0" to make it perfectly clear what your objective was.
But anyway, it sounds like a personal preference. I would run the code through lint or similar tool and see if it also thinks it's an unwise strategy. Personally, it reads like a coding mistake.