Lately I have been seeing quite a few people who initialize their std::strings like this:
std::string a{ "test" }; // Yes they do it with a single value
Now I wouldn't use this unless it's an array or to pass an initialization list.
But it got me curious, is there any point of doing this preferably over:
std::string a = "test";
std::string a ( "test" );
All three work without a doubt and I get the difference between the latter two.
Does it give some kind of performance improvement?
Does it give some kind of performance improvement?
No.
In the end, one std::string constructor gets called. The only difference is how the compiler goes about choosing the constructor.
Developers will probably choose a style according to usage with classes where it DOES matter, perhaps because there are initialization-list constructors, or because of narrowing conversions or explicit constructors. And then use the same style for strings just for the sake of consistency.
For std::string all three paths through the compiler end up choosing the same constructor.
