Answer 1

is it log, or using size_t that is the error?

https://code.sololearn.com/cA99GlPl4PCd/?ref=app

Answer 2

yes, it seems the compiler is implicitly converting the floating point/double type returned by log to integer.

Maybe there is something in here that can explain this behaviour:
http://www.learncpp.com/cpp-tutorial/44-implicit-type-conversion-coercion/

http://en.cppreference.com/w/cpp/language/implicit_conversion

Answer 3

nice! well, I guess we should be aware that not all compilers are created equal

Answer 4

It's C++ 17. You can check the version using the __cpluscplus macro: 201703 is the value corresponding to C++ 17: https://code.sololearn.com/ckHSMY7tWjP4/?ref=app

Answer 5

how about
size_t N = float(log(64)) / float(log(4));
or
size_t N = log(64.0) / log(4.0);

Answer 6

Have you lodged a bug report?

Answer 7

It should return a double or float, but the template I suggested was a possible way to enforce that is was doing so.

Prototype(s) for c+11 are: (plural)
double log (double x);
float log (float x);
long double log (long double x);
double log (T x);

http://www.cplusplus.com/reference/cmath/log/

Answer 8

I assure you, they do read them.

Answer 9

;) Are we discussing schematics now :), we could change the float conversion to a double conversion. I am not saying this is a fix, it's a work around. I agree there is a problem that needs addressing.

As far as I am aware gcc 4.8.1 supports c++11, so yeah. there is a bug somewhere

Answer 10

lol, not even a work around, the behaviour still occurs even after forcing the double via a template. So yeah, bug problems in the assignment operation or something.

Answer 11

:) I would say you are correct! (narrowing) It would be nice if the compiler had warnings enabled, it would probably allow us to further debug this

Answer 12

yeah, I know.. but it's a work around until it actually is updated

template<typename T> float log(T num) { return float(std::log(num));}

Maybe?

Answer 13

I would lodge a bug report regarding this, as it would lend more support for a case for upgrading the compiler.

Answer 14

I updated the code I used above if you want to see. It seems it is definitely in the assignment operation.

Answer 15

Hey guys,
Seems like the compiler got updated! Check it out, your code samples work normally.

Answer 16

i think i see one error on the code in the challenges:

#include <iostream>
#define func (x) x*x

using namespace std;

cout <<func (5+2)
THE Answer is supose to be: 17
but THE Answer on THE question is another