What is the purpose of various type initialization in C++ and what is the correct one?
int i1 = 1;
int i2(1);
int i3{};
int i4 = {1};
int i5 = int{1};
-
3Who claimed that there was a unique correct choice?Kerrek SB– Kerrek SB2016-05-11 09:09:01 +00:00Commented May 11, 2016 at 9:09
-
1If you want the value to be 0, option 3 is the correct one.molbdnilo– molbdnilo2016-05-11 09:19:57 +00:00Commented May 11, 2016 at 9:19
Add a comment
|
2 Answers
int i1 = 1;
This is good old C style. Still works fine and is conventional in C++.
int i2(1);
This is C++ style. It came about because some types need more than one argument to their constructor.
int i3{};
C++11 style.
int i4 = {1};
This is not conventional.
int i5 = int{1};
This is not conventional. But it's supported due to the new "uniform initialization syntax" of C++11.
int i6 = {};
You didn't ask about this one, but it's also valid in C++11.
int i7{1};
Another bonus, this is probably the most conventional use of uniform initialization syntax in C++11.
auto i8 = int{1};
Thanks to KerrekSB for this abomination, which he credits to none other than Herb Sutter. This will probably win you friends from the "No True Modern C++" camp, and alienate your coworkers who were perfectly happy with the first syntax.
12 Comments
Kerrek SB
You forgot
auto i8 = int{1}; -- Sutter style.
John Zwinck
@KerrekSB: Great, added that!
John Zwinck
@FrankPuffer: Arguably the original C++ way of
int i2(1) was a mistake and should always have been int i2{1} which is now known as Uniform Initialization in C++11. As for why there are so many options, it's because of this: xkcd.com/927 . The old way reminds people of C and doesn't work for certain types; the newest ways don't work with old compilers and will scare some people off (the majority of C++ programmers have probably never used C++11 Uniform Initialization).
TartanLlama
Note that there are situations in which
T t (1) and T t {1} will do different things. std::vector, for instance. |
TL;DR - Use X a1 {v}; for initialization.
To take cue from Bjarne Stroustrup himself, below is directly quoted from TCPL, fourth edition -
If an initializer is specified for an object, that initializer determines the initial value of an object. An initializer can use one of four syntactic styles:
X a1 {v};
X a2 = {v};
X a3 = v;
X a4(v);
Of these, only the first can be used in every context, and I strongly recommend its use. It is clearer and less error-prone than the alternatives. However, the first form (used for a1) is new in C++11, so the other three forms are what you find in older code. The two forms using = are what you use in C.
2 Comments
Bo Persson
Unfortunately the uniform initialization is not fully uniform.
std::vector<int> v1(1, 2);and std::vector<int> v2{1, 2}; will do different things.
quantumbit
So if I have s type with 2 params in constructor, the
type(1, 3) calls constructor of type with 2 parameters like a function call: ctor(1, 3), but what type{1, 2} does?