The new keyword auto; When should it be used to declare a variable type?
I think when the type is very well-known amongst the co-programmers who work (or would work) in your project, then auto
can be used, such as in the following code:
//good : auto increases readability here
for(auto it = v.begin(); it != v.end(); ++it) //v is some [std] container
{
//..
}
Or, more generally,
//good : auto increases readability here
for(auto it = std::begin(v); it != std::end(v); ++it)//v could be array as well
{
//..
}
But when the type is not very well-known and infrequently used , then I think auto
seems to reduce readability, such as here:
//bad : auto decreases readability here
auto obj = ProcessData(someVariables);
While in the former case, the usage of auto
seems very good and doesn't reduce readability, and therefore, can be used extensively, but in the latter case, it reduces readabilty and hence shouldn't be used.
Another place where auto
can be used is when you use new
1 or make_*
functions , such as here:
//without auto. Not that good, looks cumbersome
SomeType<OtherType>::SomeOtherType * obj1 = new SomeType<OtherType>::SomeOtherType();
std::shared_ptr<XyzType> obj2 = std::make_shared<XyzType>(args...);
std::unique_ptr<XyzType> obj2 = std::make_unique<XyzType>(args...);
//With auto. good : auto increases readability here
auto obj1 = new SomeType<OtherType>::SomeOtherType();
auto obj2 = std::make_shared<XyzType>(args...);
auto obj3 = std::make_unique<XyzType>(args...);
Here it is very good, as it reduces the use of keyboard, without reducing the readability, as anyone can know the type of objects being created, just by looking at the code.
1. Avoid using new
and raw-pointers though.
Sometime, the type is so irrelevant that the knowledge of the type is not even needed, such as in expression template; in fact, practically it is impossible to write the type (correctly), in such cases auto
is a relief for programmers. I've written expression template library which can be used as:
foam::composition::expression<int> x;
auto s = x * x; //square
auto c = x * x * x; //cube
for(int i = 0; i < 5 ; i++ )
std::cout << s(i) << ", " << c(i) << std::endl;
Output:
0, 0
1, 1
4, 8
9, 27
16, 64
Now compare the above code with the following equivalent code which doesn't use auto
:
foam::composition::expression<int> x;
//scroll horizontally to see the complete type!!
foam::composition::expression<foam::composition::details::binary_expression<foam::composition::expression<int>, foam::composition::expression<int>, foam::operators::multiply>> s = x * x; //square
foam::composition::expression<foam::composition::details::binary_expression<foam::composition::expression<foam::composition::details::binary_expression<foam::composition::expression<int>, foam::composition::expression<int>, foam::operators::multiply> >, foam::composition::expression<int>, foam::operators::multiply>> c = x * x * x; //cube
for(int i = 0; i < 5 ; i++ )
std::cout << s(i) << ", " << c(i) << std::endl;
As you can see, in such cases auto
makes your life exponentially easier. The expressions used above are very simple; think about the type of some more complex expressions:
auto a = x * x - 4 * x + 4;
auto b = x * (x + 10) / ( x * x+ 12 );
auto c = (x ^ 4 + x ^ 3 + x ^ 2 + x + 100 ) / ( x ^ 2 + 10 );
The type of such expressions would be even more huge and ugly, but thanks to auto
, we now can let the compiler infer the type of the expressions.
So the bottomline is: the keyword auto
might increase or decrease clarity and readability of your code, depending on the context. If the context makes it clear what type it is, or at least how it should be used (in case of standard container iterator) or the knowledge of the actual type is not even needed (such as in expression templates), then auto
should be used, and if the context doesn't make it clear and isn't very common (such as the second case above), then it should better be avoided.
Why do I need to explicitly write the 'auto' keyword?
Dropping the explicit auto
would break the language:
e.g.
int main()
{
int n;
{
auto n = 0; // this shadows the outer n.
}
}
where you can see that dropping the auto
would not shadow the outer n
.
Is there any reason to use the 'auto' keyword in C++03?
auto
is a storage class specifier, static
, register
and extern
too. You can only use one of these four in a declaration.
Local variables (without static
) have automatic storage duration, which means they live from the start of their definition until the end of their block. Putting auto in front of them is redundant since that is the default anyway.
I don't know of any reason to use it in C++. In old C versions that have the implicit int rule, you could use it to declare a variable, like in:
int main(void) { auto i = 1; }
To make it valid syntax or disambiguate from an assignment expression in case i
is in scope. But this doesn't work in C++ anyway (you have to specify a type). Funny enough, the C++ Standard writes:
An object declared without a storage-class-specifier at block scope or declared as a function parameter has automatic storage duration by default. [Note: hence, the auto specifier is almost always redundant and not often used; one use of auto is to distinguish a declaration-statement from an expression-statement (6.8) explicitly. — end note]
which refers to the following scenario, which could be either a cast of a
to int
or the declaration of a variable a
of type int
having redundant parentheses around a
. It is always taken to be a declaration, so auto
wouldn't add anything useful here, but would for the human, instead. But then again, the human would be better off removing the redundant parentheses around a
, I would say:
int(a);
With the new meaning of auto
arriving with C++0x, I would discourage using it with C++03's meaning in code.
How much is too much with C++11 auto keyword?
I think that one should use the auto
keyword whenever it's hard to say how to write the type at first sight, but the type of the right hand side of an expression is obvious. For example, using:
my_multi_type::nth_index<2>::type::key_type::composite_key_type::
key_extractor_tuple::tail_type::head_type::result_type
to get the composite key type in boost::multi_index
, even though you know that it is int
. You can't just write int
because it could be changed in the future. I would write auto
in this case.
So if the auto
keyword improves readability in a particular case then use it. You can write auto
when it is obvious to the reader what type auto
represents.
Here are some examples:
auto foo = std::make_shared<Foo>(); // obvious
auto foo = bla(); // unclear. don't know which type `foo` has
const size_t max_size = 100;
for ( auto x = max_size; x > 0; --x ) // unclear. could lead to the errors
// since max_size is unsigned
std::vector<some_class> v;
for ( auto it = v.begin(); it != v.end(); ++it )
// ok, since I know that `it` has an iterator type
// (don't really care which one in this context)
How does the this keyword work, and when should it be used?
this
is a keyword in JavaScript that is a property of an execution context. Its main use is in functions and constructors.
The rules for this
are quite simple (if you stick to best practices).
Technical description of this
in the specification
The ECMAScript standard defines this
via the abstract operation (abbreviated AO) ResolveThisBinding:
The [AO] ResolveThisBinding […] determines the binding of the keyword
this
using the LexicalEnvironment of the running execution context. [Steps]:
- Let envRec be GetThisEnvironment().
- Return ? envRec.GetThisBinding().
Global Environment Records, module Environment Records, and function Environment Records each have their own GetThisBinding method.
The GetThisEnvironment AO finds the current running execution context’s LexicalEnvironment and finds the closest ascendant Environment Record (by iteratively accessing their [[OuterEnv]] properties) which has a this binding (i.e. HasThisBinding returns true). This process ends in one of the three Environment Record types.
The value of this
often depends on whether code is in strict mode.
The return value of GetThisBinding reflects the value of this
of the current execution context, so whenever a new execution context is established, this
resolves to a distinct value. This can also happen when the current execution context is modified. The following subsections list the five cases where this can happen.
You can put the code samples in the AST explorer to follow along with specification details.
1. Global execution context in scripts
This is script code evaluated at the top level, e.g. directly inside a <script>
:
<script>
// Global context
console.log(this); // Logs global object.
setTimeout(function(){
console.log("Not global context");
});
</script>
When in the initial global execution context of a script, evaluating this
causes GetThisBinding to take the following steps:
The GetThisBinding concrete method of a global Environment Record envRec […] [does this]:
- Return envRec.[[GlobalThisValue]].
The [[GlobalThisValue]] property of a global Environment Record is always set to the host-defined global object, which is reachable via globalThis
(window
on Web, global
on Node.js; Docs on MDN). Follow the steps of InitializeHostDefinedRealm to learn how the [[GlobalThisValue]] property comes to be.
2. Global execution context in modules
Modules have been introduced in ECMAScript 2015.
This applies to modules, e.g. when directly inside a <script type="module">
, as opposed to a simple <script>
.
When in the initial global execution context of a module, evaluating this
causes GetThisBinding to take the following steps:
The GetThisBinding concrete method of a module Environment Record […] [does this]:
- Return undefined.
In modules, the value of this
is always undefined
in the global context. Modules are implicitly in strict mode.
3. Entering eval code
There are two kinds of eval
calls: direct and indirect. This distinction exists since the ECMAScript 5th edition.
- A direct
eval
call usually looks likeeval(
…);
or(eval)(
…);
(or((eval))(
…);
, etc.).1 It’s only direct if the call expression fits a narrow pattern.2 - An indirect
eval
call involves calling the function referenceeval
in any other way. It could beeval?.(
…)
,(
…, eval)(
…)
,window.eval(
…)
,eval.call(
…,
…)
, etc. Givenconst aliasEval1 = eval; window.aliasEval2 = eval;
, it would also bealiasEval1(
…)
,aliasEval2(
…)
. Separately, givenconst originalEval = eval; window.eval = (x) => originalEval(x);
, callingeval(
…)
would also be indirect.
See chuckj’s answer to “(1, eval)('this') vs eval('this') in JavaScript?” and Dmitry Soshnikov’s ECMA-262-5 in detail – Chapter 2: Strict Mode (archived) for when you might use an indirect eval()
call.
PerformEval executes the eval
code. It creates a new declarative Environment Record as its LexicalEnvironment, which is where GetThisEnvironment gets the this
value from.
Then, if this
appears in eval
code, the GetThisBinding method of the Environment Record found by GetThisEnvironment is called and its value returned.
And the created declarative Environment Record depends on whether the eval
call was direct or indirect:
- In a direct eval, it will be based on the current running execution context’s LexicalEnvironment.
- In an indirect eval, it will be based on the [[GlobalEnv]] property (a global Environment Record) of the Realm Record which executed the indirect eval.
Which means:
- In a direct eval, the
this
value doesn’t change; it’s taken from the lexical scope that calledeval
. - In an indirect eval, the
this
value is the global object (globalThis
).
What about new Function
? — new Function
is similar to eval
, but it doesn’t call the code immediately; it creates a function. A this binding doesn’t apply anywhere here, except when the function is called, which works normally, as explained in the next subsection.
4. Entering function code
Entering function code occurs when calling a function.
There are four categories of syntax to invoke a function.
- The EvaluateCall AO is performed for these three:3
- Normal function calls
- Optional chaining calls
- Tagged templates
- And EvaluateNew is performed for this one:3
- Constructor invocations
The actual function call happens at the Call AO, which is called with a thisValue determined from context; this argument is passed along in a long chain of call-related calls. Call calls the [[Call]] internal slot of the function. This calls PrepareForOrdinaryCall where a new function Environment Record is created:
A function Environment Record is a declarative Environment Record that is used to represent the top-level scope of a function and, if the function is not an ArrowFunction, provides a
this
binding. If a function is not an ArrowFunction function and referencessuper
, its function Environment Record also contains the state that is used to performsuper
method invocations from within the function.
In addition, there is the [[ThisValue]] field in a function Environment Record:
This is the
this
value used for this invocation of the function.
The NewFunctionEnvironment call also sets the function environment’s [[ThisBindingStatus]] property.
[[Call]] also calls OrdinaryCallBindThis, where the appropriate thisArgument is determined based on:
- the original reference,
- the kind of the function, and
- whether or not the code is in strict mode.
Once determined, a final call to the BindThisValue method of the newly created function Environment Record actually sets the [[ThisValue]] field to the thisArgument.
Finally, this very field is where a function Environment Record’s GetThisBinding AO gets the value for this
from:
The GetThisBinding concrete method of a function Environment Record envRec […] [does this]:
[…]
3. Return envRec.[[ThisValue]].
Again, how exactly the this value is determined depends on many factors; this was just a general overview. With this technical background, let’s examine all the concrete examples.
Arrow functions
When an arrow function is evaluated, the [[ThisMode]] internal slot of the function object is set to “lexical” in OrdinaryFunctionCreate.
At OrdinaryCallBindThis, which takes a function F:
- Let thisMode be F.[[ThisMode]].
- If thisMode is lexical, return NormalCompletion(
undefined
).
[…]
which just means that the rest of the algorithm which binds this is skipped. An arrow function does not bind its own this value.
So, what is this
inside an arrow function, then? Looking back at ResolveThisBinding and GetThisEnvironment, the HasThisBinding method explicitly returns false.
The HasThisBinding concrete method of a function Environment Record envRec […] [does this]:
- If envRec.[[ThisBindingStatus]] is lexical, return false; otherwise, return true.
So the outer environment is looked up instead, iteratively. The process will end in one of the three environments that have a this binding.
This just means that, in arrow function bodies, this
comes from the lexical scope of the arrow function, or in other words (from Arrow function vs function declaration / expressions: Are they equivalent / exchangeable?):
Arrow functions don’t have their own
this
[…] binding. Instead, [this identifier is] resolved in the lexical scope like any other variable. That means that inside an arrow function,this
[refers] to the [value ofthis
] in the environment the arrow function is defined in (i.e. “outside” the arrow function).
Function properties
In normal functions (function
, methods), this
is determined by how the function is called.
This is where these “syntax variants” come in handy.
Consider this object containing a function:
const refObj = {
func: function(){
console.log(this);
}
};
Alternatively:
const refObj = {
func(){
console.log(this);
}
};
In any of the following function calls, the this
value inside func
will be refObj
.1
refObj.func()
refObj["func"]()
refObj?.func()
refObj.func?.()
refObj.func``
If the called function is syntactically a property of a base object, then this base will be the “reference” of the call, which, in usual cases, will be the value of this
. This is explained by the evaluation steps linked above; for example, in refObj.func()
(or refObj["func"]()
), the CallMemberExpression is the entire expression refObj.func()
, which consists of the MemberExpression refObj.func
and the Arguments ()
.
But also, refObj.func
and refObj
play three roles, each:
- they’re both expressions,
- they’re both references, and
- they’re both values.
refObj.func
as a value is the callable function object; the corresponding reference is used to determine the this
binding.
The optional chaining and tagged template examples work very similarly: basically, the reference is everything before the ?.()
, before the ``
, or before the ()
.
EvaluateCall uses IsPropertyReference of that reference to determine if it is a property of an object, syntactically. It’s trying to get the [[Base]] property of the reference (which is e.g. refObj
, when applied to refObj.func
; or foo.bar
when applied to foo.bar.baz
). If it is written as a property, then GetThisValue will get this [[Base]] property and use it as the this value.
Note: Getters / Setters work the same way as methods, regarding this
. Simple properties don’t affect the execution context, e.g. here, this
is in global scope:
const o = {
a: 1,
b: this.a, // Is `globalThis.a`.
[this.a]: 2 // Refers to `globalThis.a`.
};
Calls without base reference, strict mode, and with
A call without a base reference is usually a function that isn’t called as a property. For example:
func(); // As opposed to `refObj.func();`.
This also happens when passing or assigning methods, or using the comma operator. This is where the difference between Reference Record and Value is relevant.
Note function j
: following the specification, you will notice that j
can only return the function object (Value) itself, but not a Reference Record. Therefore the base reference refObj
is lost.
const g = (f) => f(); // No base ref.
const h = refObj.func;
const j = () => refObj.func;
g(refObj.func);
h(); // No base ref.
j()(); // No base ref.
(0, refObj.func)(); // Another common pattern to remove the base ref.
EvaluateCall calls Call with a thisValue of undefined here. This makes a difference in OrdinaryCallBindThis (F: the function object; thisArgument: the thisValue passed to Call):
- Let thisMode be F.[[ThisMode]].
[…]
- If thisMode is strict, let thisValue be thisArgument.
- Else,
- If thisArgument is undefined or null, then
- Let globalEnv be calleeRealm.[[GlobalEnv]].
- […]
- Let thisValue be globalEnv.[[GlobalThisValue]].
- Else,
- Let thisValue be ! ToObject(thisArgument).
- NOTE: ToObject produces wrapper objects […].
[…]
Note: step 5 sets the actual value of this
to the supplied thisArgument in strict mode — undefined
in this case. In “sloppy mode”, an undefined or null thisArgument results in this
being the global this value.
If IsPropertyReference returns false, then EvaluateCall takes these steps:
- Let refEnv be ref.[[Base]].
- Assert: refEnv is an Environment Record.
- Let thisValue be refEnv.WithBaseObject().
This is where an undefined thisValue may come from: refEnv.WithBaseObject() is always undefined, except in with
statements. In this case, thisValue will be the binding object.
There’s also Symbol.unscopables
(Docs on MDN) to control the with
binding behavior.
To summarize, so far:
function f1(){
console.log(this);
}
function f2(){
console.log(this);
}
function f3(){
console.log(this);
}
const o = {
f1,
f2,
[Symbol.unscopables]: {
f2: true
}
};
f1(); // Logs `globalThis`.
with(o){
f1(); // Logs `o`.
f2(); // `f2` is unscopable, so this logs `globalThis`.
f3(); // `f3` is not on `o`, so this logs `globalThis`.
}
and:
"use strict";
function f(){
console.log(this);
}
f(); // Logs `undefined`.
// `with` statements are not allowed in strict-mode code.
Note that when evaluating this
, it doesn’t matter where a normal function is defined.
.call
, .apply
, .bind
, thisArg, and primitives
Another consequence of step 5 of OrdinaryCallBindThis, in conjunction with step 6.2 (6.b in the spec), is that a primitive this value is coerced to an object only in “sloppy” mode.
To examine this, let’s introduce another source for the this value: the three methods that override the this binding:4
Function.prototype.apply(thisArg, argArray)
Function.prototype.
{call
,bind
}(thisArg, ...args)
.bind
creates a bound function, whose this binding is set to thisArg and cannot change again. .call
and .apply
call the function immediately, with the this binding set to thisArg.
.call
and .apply
map directly to Call, using the specified thisArg. .bind
creates a bound function with BoundFunctionCreate. These have their own [[Call]] method which looks up the function object’s [[BoundThis]] internal slot.
Examples of setting a custom this value:
function f(){
console.log(this);
}
const myObj = {},
g = f.bind(myObj),
h = (m) => m();
// All of these log `myObj`.
g();
f.bind(myObj)();
f.call(myObj);
h(g);
For objects, this is the same in strict and non-strict mode.
Now, try to supply a primitive value:
function f(){
console.log(this);
}
const myString = "s",
g = f.bind(myString);
g(); // Logs `String { "s" }`.
f.call(myString); // Logs `String { "s" }`.
In non-strict mode, primitives are coerced to their object-wrapped form. It’s the same kind of object you get when calling Object("s")
or new String("s")
. In strict mode, you can use primitives:
"use strict";
function f(){
console.log(this);
}
const myString = "s",
g = f.bind(myString);
g(); // Logs `"s"`.
f.call(myString); // Logs `"s"`.
Libraries make use of these methods, e.g. jQuery sets the this
to the DOM element selected here:
$("button").click(function(){
console.log(this); // Logs the clicked button.
});
Constructors, classes, and new
When calling a function as a constructor using the new
operator, EvaluateNew calls Construct, which calls the [[Construct]] method. If the function is a base constructor (i.e. not a class extends
…{
…}
), it sets thisArgument to a new object created from the constructor’s prototype. Properties set on this
in the constructor will end up on the resulting instance object. this
is implicitly returned, unless you explicitly return your own non-primitive value.
A class
is a new way of creating constructor functions, introduced in ECMAScript 2015.
function Old(a){
this.p = a;
}
const o = new Old(1);
console.log(o); // Logs `Old { p: 1 }`.
class New{
constructor(a){
this.p = a;
}
}
const n = new New(1);
console.log(n); // Logs `New { p: 1 }`.
Class definitions are implicitly in strict mode:
class A{
m1(){
return this;
}
m2(){
const m1 = this.m1;
console.log(m1());
}
}
new A().m2(); // Logs `undefined`.
super
The exception to the behavior with new
is class extends
…{
…}
, as mentioned above. Derived classes do not immediately set their this value upon invocation; they only do so once the base class is reached through a series of super
calls (happens implicitly without an own constructor
). Using this
before calling super
is not allowed.
Calling super
calls the super constructor with the this value of the lexical scope (the function Environment Record) of the call. GetThisValue has a special rule for super
calls. It uses BindThisValue to set this
to that Environment Record.
class DerivedNew extends New{
constructor(a, a2){
// Using `this` before `super` results in a ReferenceError.
super(a);
this.p2 = a2;
}
}
const n2 = new DerivedNew(1, 2);
console.log(n2); // Logs `DerivedNew { p: 1, p2: 2 }`.
5. Evaluating class fields
Instance fields and static fields were introduced in ECMAScript 2022.
When a class
is evaluated, ClassDefinitionEvaluation is performed, modifying the running execution context. For each ClassElement:
- if a field is static, then
this
refers to the class itself, - if a field is not static, then
this
refers to the instance.
Private fields (e.g. #x
) and methods are added to a PrivateEnvironment.
Static blocks are currently a TC39 stage 3 proposal. Static blocks work the same as static fields and methods: this
inside them refers to the class itself.
Note that in methods and getters / setters, this
works just like in normal function properties.
Related Topics
Is the "One-Past-The-End" Pointer of a Non-Array Type a Valid Concept in C++
Using the Less Than Comparison Operator for Strings
C++ "Error: Passing 'Const Std::Map<Int, Std::Basic_String<Char> >' as 'This' Argument of ..."
Why Are Empty Expressions Legal in C/C++
Global Function Definition in Header File - How to Avoid Duplicated Symbol Linkage Error
Cyclic Dependency Between Header Files
Rotate an Image in C++ Without Using Opencv Functions
G++ Variable Size Array No Warning
Why Was the Space Character Not Chosen for C++14 Digit Separators
C++ Global Variable Declaration
If I Do a 'Typedef' in C or C++, When Should I Add '_T' at the End of Typedef'Ed Type
Why Are C++ Stl iOStreams Not "Exception Friendly"
How to Add Constructors/Destructors to an Unnamed Class
Macro Definition Containing #Include Directive
C++/Win32: How to Wait for a Pending Delete to Complete
Forwarding All Constructors in C++0X