Chaining Bool values give opposite result to expected - c++

Unthinkingly I wrote some code to check that all the values of a struct were set to 0. To accomplish this I used:
bool IsValid() {
return !(0 == year == month == day == hour == minute == second);
}
where all struct members were of type unsigned short. I used the code as part of a larger test but noticed that it was returning false for values differing from zero, and true for values that were all equal to zero - the opposite of what I expected.
I changed the code to read:
bool IsValid() {
return (0 != year) || (0 != month) || (0 != day) || (0 != hour) || (0 != minute) || (0 != second);
}
But would like to know what caused the odd behaviour. Is it a result of precedence? I've tried to Google this answer but found nothing, if there's any nomenclature to describe the result I'd love to know it.
I compiled the code using VS9 and VS8.

== groups from left to right, so if all values are zero then:
0 == year // true
(0 == year) == month // false, since month is 0 and (0 == year) converts to 1
((0 == year) == month) == day // true
And so on.
In general, x == y == z is not equivalent to x == y && x == z as you seem to expect.

The behaviour shouldn't be seen as odd. The grammar rules for == (and most but not all binary operators) specify left to right grouping so your original expression is equivalent to:
!((((((0 == year) == month) == day) == hour) == minute) == second)
Note that when compared to an integer type a bool expression with value true will promote to 1 and with value false will promote to 0. (In C the result of the equality operator is an int in any case with a value or either 1 or 0.)
This means that, for example, ((0 == year) == month) will be true if year is zero and month is one or if year is non-zero but month is zero and false otherwise.

You have to consider how it's evaluated...
a == b == c
is asking if two of them are equal (a and b), then comparing that boolean result to the third value c! It is NOT comparing the first two values with the third. Anything beyond 2 arguments won't chain as you evidently expect.
For whatever it's worth, because C++ considers non-0 values to be "true" in a boolean context, you can express what you want simply as:
return year && month && day && hour && minute && second;
(note: your revised code says "month" twice and doesn't test minute).
Back to the chained ==s: with user-defined types and operator overloading you can create a class that compares as you expect (and it can even allow things like 0 <= x < 10 to "work" in the way it's read in mathematics), but creating something special will just confuse other programmers who already know the (weird) way these things work for builtin types in C++. Worth doing as a ten/twenty minute programming exercise though if you're keen to learn C++ in depth (hint: you need the comparison operators to return a proxy object that remembers what will be the left-hand-side value for the next comparison operator).
Finally, sometimes these "weird" boolean expressions are useful: for example, a == b == (c == d) might be phrased in English as "either (a == b) and (c == d), OR (a != b) and (c != d)", or perhaps "the equivalence of a and b is the same as the equivalence of c and d (whether true or false doesn't matter)". That might model real world situations like a double-dating scenario: if a likes/dislikes b (their date) as much as c likes/dislikes d, then they'll either hang around and have a nice time or call it quits quickly and it's painless either way... otherwise one couple will have a very tedious time of it.... Because these things can make sense, it's impossible for the compiler to know you didn't intend to create such an expression.

Your error here is writing a mathematical expression using equals signs, and unthinkingly supposing that the computer will perform the test you meant - what a human mathematician would see as the meaning of those symbols. What the computer does (as per the definition of the language) is to perform a series of discrete comparisons, each of which returns true or false - and this true or false is then used in the next comparison. You aren't comparing all of those variables to 0, you're comparing each (bar two of them) to the result of comparing another two of the said variables.

The return of the == operator is 1 if the operands are equal, so regardless wether this is read from left to right or right to left, this will not do what you expect.
so this could only work in an analogous test if you would be interested if all values are 1.
And to have a shorter expression since you seem interested in that just do year || day || ...

Related

Find a logical expression that is true if `n` is a multiple of `2019` and is not in the interval `(a, b)`

I had the task of finding a logical expression that would result in 1 if and only if a given number n is a multiple of 2019 and is NOT from the interval (a, b).
The textbook gave the following answer and I don't really understand it:
a>=n || b<=n && (n%3==0 && n%673==0)
The thing between those parantheses I understand to be equivalent to n%2019==0, so that's alright. But I don't understand why this works, I mean the && operator has higher priority that the || operator, so wouldn't we evaluate
b<=n && (n%3==0 && n%673==0)
first and only at the end if n<=a? I thought that if I were to do it, I would do it like this:
(a>=n || b<=n) && (n%3==0 && n%673==0)
So I just added that extra set of parantheses. Now we would check if the number is not in the interval (a, b), then we would check if it is a multiple of 2019 and then we would 'and' those to answers to get the final answer. This makes sense to me. But I don't understand why they omitted that set of parantheses, why would that still work? Shouldn't we consider that && has higher priority than ||, so we add an extra set of parantheses? Would it still work? Or is it me that is wrong?
Trying it out shows that the expression as written without the extra parentheses doesn't work:
bool expr(int n, int a, int b)
{
return a>=n || b<=n && (n%3==0 && n%673==0);
}
expr(1000, 2000, 2018) for example evaluates to true, even though it is not a multiple of 2019.
As you pointed out, the logical AND operator && has higher precedence than the logical OR operator || (reference), so the expression is equivalent to:
a>=n || (b<=n && (n%3==0 && n%673==0))
which is always true when n <= a, even if it's not a multiple of 2019.
A clearer expression would be:
(n % 2019 == 0) && (n <= a || n >= b)

Integer to Boolean strange syntax [duplicate]

This question already has an answer here:
The Definitive C++ Book Guide and List
(1 answer)
Closed 7 years ago.
I'm less than a year into C++ development (focused on other languages prior to this) and I'm looking at a guy's code who's been doing this for two decades. I've never seen this syntax before and hopefully someone can be of some help.
bool b; // There exists a Boolean variable.
int i; // There exists an integer variable.
sscanf(value, "%d", &i); // The int is assigned from a scan.
b = (i != 0); // I have never seen this syntax before.
I get that the boolean is being assigned from the int that was just scanned, but I don't get the (* != 0) aspects of what's going on. Could someone explain why this person who knows the language much better than I is doing syntax like this?
Have a read here:
http://en.cppreference.com/w/cpp/language/operator_comparison
The result of operator != is a bool. So the person is saying "compare the value in i with 0". If 'i' is not equal to 0, then the '!=' returns true.
So in effect the value in b is "true if 'i' is anything but zero"
EDIT: In response to the OP's comment on this, yes you could have a similar situation if you used any other operator which returns bool. Of course when used with an int type, the != means negative numbers evaluate to true. If > 0 were used then both 0 and negative numbers would evaluate to false.
The expression (i != 0) evaluates to a boolean value, true if the expression is true (i.e. if i is non-zero) and false otherwise.
This value is then assigned to b.
You'd get the same result from b = i;, if you prefer brevity to explicitness, due to the standard boolean conversion from numeric types which gives false for zero and true for non-zero.
Or b = (i != 0) ? true : false; if you like extraneous verbosity.
(i != 0) is an expression that evaluates to true or false. Hence, b gets the value of true/false depending on the value of i.
This is fairly fundamental syntax. The != operator performs a "not equal to" comparison.
You may be being confused by the shorthand of initialising a bool directly from the result of a comparison operator, but the syntax itself is not esoteric.
The program is essentially equivalent to:
bool b;
int i;
sscanf(value, "%d", &i);
if (i != 0)
b = true;
else
b = false;
The key is that i != 0 is itself an expression that evaluates to true or false, not some magic that may only be used in an if statement.
Basically, if the condition (i not_equal_to 0 ) is satisfied, b gets the value "true". Else b gets the value "false".
Here, "i != 0" is a boolean expression that will be true if "i" is non-zero and false if it is zero.
All that is happening here is the result of that expression is being assigned to a variable.
You could also do things like...
boolean canDrinkAlcohol = (person.age() >= 18 && person.country.equals("UK") || person.age() >= 21 && person.county.equals("US"));
...
if(canDrinkAlcohol) {
...
}
or something

Comparison of int values in an if statement

I had a question in my test paper in which we had to compare the values of int type variables. The first thought that came to my mind was that it was missing the && operator but i am not sure.
int a=2, b=2, c=2;
if(a==b==c)
{
printf("hello");
}
I have a doubt, will the above statement will execute or not in c or c++? Can i have the reason as well.
Thank You
It will execute but with what I believe unexpected results to you.
One of the == will evaluate to a boolean value, which will then be converted to an int and then the second comparison will be performed, comparing an int to either 1 or 0.
The correct statement is a==b && b==c.
For example:
3 == 3 == 3
evaluates to
true == 3
1 == 3
false
a==b==c
is equivalent to
(a == b) == c
The result of a == b is 1 (if true) or 0 (if false), so it will probably not achieve what you expect.
Use a == b && b == c to check if the value of the three objects are equal.
a == b == c is a comparison between c and result of a==b (1 or 0) operation.
use a==b&&b==c.
the condition a==b==c is equivalent to (a==b)==c which will provide the required result iff c==1, else the code will fail.

C++ bitwise OR operator

bool OrderUtils::shouldCptyAutoexecute(int Id)
{
bool res =
dummyCache::instance().getVal(Id).getWhitelabelType() == "ABCD";
if (!res)
res |= dummyCache::instance().getVal(Id).getIsPlc() == 1;
return res;
}
The above code checks for 2 Id's and returns true to res if any of the id is present in the database.
Can you suggest a way in which I can compare one more value from the databse table and return true to the value res..Also can you explain what does the second if statement do and the bitwise OR operator?
Sir, just let the short-circuit eval do this for you:
return dummyCache::instance().getVal(Id).getWhitelabelType() == "ABCD" ||
dummyCache::instance().getVal(Id).getIsPlc() == 1;
If the first is true, the second will not fire. Moreover, I assure you a remotely-reasonable optimizing compiler will not re-fire instance().getVal(id) if the returned object has not changed between the getWhitelabelType() and getisPlc() calls. In fact, i would all-but-guarantee it if getWhiteLabelType() is const. (and it certainly looks like it should be).
Regarding the bit work. The expression was pretty-much whacked. though it will work. Unless I read it wrong (and trust me, the list of people that will tell me I am will let me know quickly) it is performing a boolean eval, promoting the resulting true/false bool to an int, promoting the current value of res from bool to int (which is zero, so nothing special there), bitwise-OR-ing that with the expression int, then demoting the final int back to a bool to store in res . I'm surprised this doesn't flag at least a warning on the compiler.
It likely should have been if (!res) res ||= expr, and even then, it is pointless, as you can just use short circuit eval as in the top of this answer to cut out the local res entirely. Consider if res were false. Then the equivalent expression would be res = false || expr. But thats just res = expr. In the !res state it executes in, you may as well just use an assignment.
Finally, regarding adding a third field to your eval, it depends entirely on how you want it added. for an additional logical OR it is pretty simple.
For an expression like (A || B || C) you can just
return dummyCache::instance().getVal(Id).AField() == ATestValue ||
dummyCache::instance().getVal(Id).BField() == BTestValue ||
dummyCache::instance().getVal(Id).CField() == CTestValue;
For more complex operations, some judicious use of parenthesis will go a long way. For example, to return (A || B) && C:
return (dummyCache::instance().getVal(Id).AField() == ATestValue ||
dummyCache::instance().getVal(Id).BField() == BTestValue) &&
dummyCache::instance().getVal(Id).CField() == CTestValue;
Or perhaps (A && C) || (B && !C) (ok this is getting a little overboard...)
return (dummyCache::instance().getVal(Id).CField() == CTestValue)
? (dummyCache::instance().getVal(Id).AField() == ATestValue)
: (dummyCache::instance().getVal(Id).BField() == BTestValue);

Is it possible to define a variable in expression in C++?

I have this insane homework where I have to create an expression to validate date with respect to Julian and Gregorian calendar and many other things ...
The problem is that it must be all in one expression, so I can't use any ;
Are there any options of defining variable in expression? Something like
d < 31 && (bool leapyear = y % 4 == 0) || (leapyear ? d % 2 : 3) ....
where I could define and initialize one or more variables and use them in that one expression without using any ;?
Edit: It is explicitly said, that it must be a one-line expression. No functions ..
The way I'm doing it right now is writing macros and expanding them, so I end up with stuff like this
#define isJulian(d, m, y) (y < 1751 || (y == 1752 && (m < 9) || (m == 9 && d <= 2)))
#define isJulianLoopYear(y) (y % 4 == 0)
#define isGregorian(d, m, y) (y > 1573 || (y == 1752 && (m > 9) || (m == 9 && d > 13)))
#define isGregorianLoopYear(y) ((y % 4 == 0) || (y % 400 = 0))
// etc etc ....
looks like this is the only suitable way to solve the problem
edit: Here is original question
Suppose we have variables d m and y containing day, month and year. Task is to write one single expression which decides, if date is valid or not. Value should be true (non-zero value) if date is valid and false (zero) if date is not valid.
This is an example of expression (correct expression would look something like this):
d + 4 == y ^ 85 ? ~m : d * (y-2)
These are examples of wrong answers (not expressions):
if ( log ( d ) == 1752 ) m = 1;
or:
for ( i = 0; i < 32; i ++ ) m = m / 2;
Submit only file containing only one single expression without ; at the end. Don't submit commands or whole program.
Until 2.9.1752 was Julian calendar, after that date is Gregorian calendar
In Julian calendar is every year dividable by 4 a leap year.
In Gregorian calendar is leap year ever year, that is dividible by 4, but is not dividible by 100. Years that are dividable by 400 are another exception and are leap years.
1800, 1801, 1802, 1803, 1805, 1806, ....,1899, 1900, 1901, ... ,2100, ..., 2200 are not loop years.
1896, 1904, 1908, ..., 1996, 2000, 2004, ..., 2396,..., 2396, 2400 are loop years
In september 1752 is another exception, when 2.9.1752 was followed by 14.9.1752, so dates 3.9.1752, 4.9.1752, ..., 13.9.1752 are not valid.
((m >0)&&(m<13)&&(d>0)&&(d<32)&&(y!=0)&&(((d==31)&&
((m==1)||(m==3)||(m==5)||(m==7)||(m==8)||(m==10)||(m==12)))
||((d<31)&&((m!=2)||(d<29)))||((d==29)&&(m==2)&&((y<=1752)?((y%4)==0):
((((y%4)==0)&&((y%100)!=0))
||((y%400)==0)))))&&(((y==1752)&&(m==9))?((d<3)||(d>13)):true))
<evil>
Why would you define a new one, if you can reuse an existing one? errno is a perfect temporary variable.
</evil>
I think the intent of the homework is to ask you to do this without using variables, and what you are trying to do might defeat its purpose.
In standard C++, this is not possible. G++ has an extension known as statement expressions that can do that.
I don't believe you can, but even if you could, it would only have scope inside of the parentheses they are defined in (in your example) and cannot be used outside of them.
Your solution, which I will not provide fully for you, will probably go along these lines:
isJulian ? isJulianLeapyear : isGregorianLeapyear
To make it more specific, it could be like this:
isJulian ? (year % 4) == 0 : ((year % 4) == 0 || (year % 400) == 0)
You'll have to just make sure your algorithm is correct. I'm not a calender expert, so I wouldn't know about that.
First: Don't. It may be cute, but even if there's an extension that allows it, code golf is a dangerous game that will almost always end up causing mmore grief than it solves.
Okay, back to the 'real' question as defined by the homework. Can you make additional functions? If so, instead of capturing whether or not it's a leap year in a variable, make a function isLeapYear(int year) that returns the correct value.
Yes, that means you'll calculate it more than once. If that ends up being a performance issue, I'd be incredibly surprised... and it's a premature optimization to worry about that in the first place.
I'd be very surprised if you weren't allowed to write functions as part of doing this. It seems like that'd be half the point of an exercise like this.
......
Okay, so here's a quick overview of what you'll need to do.
First, basic verification - that month, day, year are possible values at all - month 0-11 (assuming 0-based), day 0-30, year non-negative (assuming that's a constraint).
Once you're past that, I'd probably check for the 1752 special cases.
If that's not relevant, the 'regular' months can be handled pretty simply.
This leaves us with the leap year cases, which can be broken down into two expressions - whether something is a leap year (which will be broken down additionally based on gregorian/julian), and whether the date is valid at that point.
So, at the highest level, your expression looks something like this:
areWithinRange(d,m,y) && passes1752SpecialCases(d,m,y) && passes30DayMonths(d,m,y) && passes31DayMonths(d,m,y) && passesFebruaryChecks(d,m,y)
If we assume that we only return false from our sub-expressions if we actively detect a rule break (31 days in June for the 30DayMonth rule returns false, but 30 days in February is irrelevant and so passes true), then we can pretty much say that the logic at that level is correct.
At this point, I'd write separate functions for the individual pieces (as pure expressions, a single return ... statement). Once you've gotten those in place, you can replace the method call in your top-level expression with the expanded version. Just make sure you parenthesize (is that a word?) everything sufficiently.
I'd also make a test harness program that uses the expression and has a number of valid and invalid inputs, and verifies that you're doing the right thing. You can write that in a function for ease of cut and paste for the final turn-in by doing something like:
bool isValidDate(int d, int m, int y)
{
return
// your expression here
}
Since the expression will be on a line by itself, it'll be easy to cut and paste.
You may find other ways to simplify your logic - excepting the 1752 special cases, days between 1 and 28 are always valid, for instance.
Considering it is homework, I think the best advice would be an approach to deriving your own solution.
If I were tackling this assignment, I would start by breaking the rules.
I would write a c++ function that given the variables d, m, and y, returns a boolean result on the validity of the date. I would use as many non-recursive helper functions as needed, and feel free to use if, else if, and else, but no looping aloud.
I would then Inline all helper functions
I would reduce all if, else if, and else statement to ? : notation
If I was successful at limiting my use of variables, I might be able to reduce all this to a single function with no variables - the body of which will contain the single expression I seek.
Good Luck.
You clearly have to have the date passed in somehow. Beyond that, all you're really going to be doing is chaining && and || (assuming we get the date as a tm struct):
#include <ctime>
bool validate(tm date)
{
return (
// sanity check that all values are positive
date.tm_mday >= 1 && date.tm_mon >= 0 && date.tm_year >= 0
// check for valid days
&& ((date.tm_mon == 0 && date.tm_mday <= 31)
|| (date.tm_mon == 1 && date.tm_mday <= (date.tm_year % 4 ? 28 : 29))
|| (date.tm_mon == 2 && date.tm_mday <= 31)
// yadda yadda for the other months
|| (date.tm_mon == 11 && date.tm_mday <= 31))
);
}
The parenthesis around date.tm_year % 4 ? 28 : 29 actually aren't needed, but I'm including them for readability.
UPDATE
Looking at a comment, you'll also need similar rules to validate dates that don't exist in the Gregorian calendar.
UPDATE II
Since you're dealing with dates in the past you will need to implement a more correct leap year test. However, I generally deal with dates in the future, and this incorrect leap year test will give correct results in 2012, 2016, 2020, 2024, 2028, 2032, 2036, 2040, 2044, 2048, 2052, 2056, 2060, 2064, 2068, 2072, 2076, 2080, 2084, 2088, 2092, and 2096. I will make a prediction that before this test fails in 2100 silicon-based computers will be forgotten relics. I seriously doubt we will use C++ on the quantum computers in use then. Besides, I won't be the programmer assigned to fix the bug.