Stanley Lippman's "C++ primer" mentioned, in page 234 that
Ordinarily, it is a bad idea to declare a function locally. However,
to explain how scope interacts with overloading, we will violate this
practice and use local function declarations.
...
void print(const string &);
void print(double); // overloads the print function
void fooBar(int ival)
{ ...
// bad practice: usually it's a bad idea to declare functions at local scope
void print(int); // new scope: hides previous instances of print
print(ival); // ok: print(int) is visible
print(3.14); // ok: calls print(int); print(double) is hidden
}
Nevertheless, which situation it might make sense to do so, declaring a function inside the body of a function?
I recall similar code in Scheme:
(define (PowerSet set)
(if (null? set) '(())
(let ((PowerSetAfterHead (PowerSet (cdr set) ) ))
(append PowerSetAfterHead
(PowerSet (cdr set) )
(map
(lambda (subset)
(cons (car set) subset
)
)
(PowerSet (cdr set) )
)
)
)
)
)
Is it normally used to hide from misuse of a "local function" that intends for the outer function's usage only? something like internal class? Or is it useful in some design patterns?
The problem arises when there is an ambiguity in calling a function. Thus to avoid the ambiguity in some scope you can redeclare the required function that hides all other functions with the same name of the outer scope.
Consider the following simple example. If you run the following program
#include <iostream>
void f( int ) { std::cout << "f( int )" << std::endl; }
void f( double ) { std::cout << "f( double )" << std::endl; }
int main()
{
f( 10l );
return 0;
}
you will get an error like
prog.cpp:10:9: error: call of overloaded 'f(long int)' is ambiguous f( 10l );
But if you will add a declaration of function f within main like this
#include <iostream>
void f( int ) { std::cout << "f( int )" << std::endl; }
void f( double ) { std::cout << "f( double )" << std::endl; }
int main()
{
void f( double );
f( 10l );
return 0;
}
the code will be compiled and the output will look like
f( double )
Your scheme example both declares and defines a local function.
The C++ example just declares a function locally.
These are different things.
You can create the rough equivalent of scheme local functions with lambdas in C++11.
The declaring of a function locally just says "there exist a function of this name and signature". What more, it does so in a narrower scope than the one outside the function, and as there are functions in said narrower scope only they are considered for overloading (not the ones in the wider scope).
No new function is created -- it must be defined elsewhere. (You cannot define a local function in C++, barring things like local class methods and lambdas)
You could use this to change how overload resolution works, but this is rarely a good idea. In the case that you have an overload ambiguity, you are better off manually converting the types and then calling the overload set, and have it resolve normally.
An alternative is to declare only the overload you want to invoke, then invoke it. This can result in surprising behavior, so do so only in a very narrow scope if you can, and document why you are doing it and what possible problems could occur (as few programmers will have the scope-hiding overload rules at the front of their brain when reading your code without at least a bit of prompting).
I do not think I have ever ran into a situation where I had to declare a function locally.
The handful of times I did it it was because I was hacking into some serious ball of mud code, and wanted to make an extremely local change for prototype/debugging purposes. I simply declared a function signature and called it somewhere. The function was a static local in another .cpp file in the same library elsewhere, I removed the static there and tested the result of calling it.
This saved creating a header file (or modifying one), exposing it formally, including said header where we where using it. Prior to committing to production I would go through steps like that (and probably clean up the interface of that static local function while I was at it).
So, I used it for a quick and dirty prototype hack. The last time I did this I ended up reverting it after I did the prototype hack, which was easier because I had only touched 2 files each at 1 spot.
Related
I was taught that functions need declarations to be called. To illustrate, the following example would give me an error as there is no declaration for the function sum:
#include <iostream>
int main() {
std::cout << "The result is " << sum(1, 2);
return 0;
}
int sum(int x, int y) {
return x + y;
}
// main.cpp:4:36: error: use of undeclared identifier 'sum'
// std::cout << "The result is " << sum(1, 2);
// ^
// 1 error generated.
To fix this, I'd add the declaration:
#include <iostream>
int sum(int x, int y); // declaration
int main() {
std::cout << "The result is " << sum(1, 2);
return 0;
}
int sum(int x, int y) {
return x + y;
}
Why the main function doesn't need the declaration, as other functions like sum need?
A definition of a function is also a declaration of a function.
The purpose of a declaring a function is to make it known to the compiler. Declaring a function without defining it allows a function to be used in places where it is inconvenient to define it. For example:
If a function is used in a source file (A) other than the one it is defined in (B), we need to declare it in A (usually via a header that A includes, such as B.h).
If two or more functions may call each other, then we cannot define all those functions before the others—one of them has to be first. So declarations can be provided first, with definitions coming afterward.
Many people prefer to put “higher level” routines earlier in a source file and subroutines later. Since those “higher level” routines call various subroutines, the subroutines must be declared earlier.
In C++, a user program never calls main, so it never needs a declaration before the definition. (Note that you could provide one if you wished. There is nothing special about a declaration of main in this regard.) In C, a program can call main. In that case, it does require that a declaration be visible before the call.
Note that main does need to be known to the code that calls it. This is special code in what is typically called the C++ runtime startup code. The linker includes that code for you automatically when you are linking a C++ program with the appropriate linker options. Whatever language that code is written in, it has whatever declaration of main it needs in order to call it properly.
I was taught that functions need declarations to be called.
Indeed. A function must be declared before it can be called.
why we don't add a declaration for the main function?
Well, you didn't call main function. In fact, you must not call main at all1, so there is never a need to declare main before anything.
Technically though, all definitions are also declarations, so your definition of main also declares main.
Footnote 1: The C++ standard says it's undefined behaviour to call main from within the program.
This allows C++ implementations to put special run-once startup code at the top of main, if they aren't able to have it run earlier from hooks in the startup code that normally calls main. Some real implementations do in fact do this, e.g. calling a fast-math function that sets some FPU flags like denormals-are-zero.
On a hypothetical implementation, calling main could result in fun things like re-running constructors for all static variables, re-initializing the data structures used by new/delete to keep track of allocations, or other total breakage of your program. Or it might not cause any problem at all. Undefined behaviour doesn't mean it has to fail on every implementation.
The prototype is required if you want to call the function, but it's not yet available, like sum in your case.
You must not call main yourself, so there is no need to have a prototype. It's even a bad a idea to write a prototype.
No, the compiler does not need a forward declaration for main().
main() is a special function in C++.
Some important things to remember about main() are:
The linker requires that one and only one main() function exist when creating an executable program.
The compiler expects a main() function in one of the following two forms:
int main () { /* body */ }
int main (int argc, char *argv[]) { /* body */ }
where body is zero or more statements
An additional acceptable form is implementation specific and provides a list of the environment variables at the time the function is called:
int main (int argc, char* argv[], char *envp[]) { /* body */ }
The coder must provide the 'definition' of main using one of these acceptable forms, but the coder does not need to provide a declaration. The coded definiton is accepted by the compiler as the declaration of main().
If no return statement is provided, the compiler will provide a return 0; as the last statement in the function body.
As an aside, there is sometimes confusion about whether a C++ program can make a call to main(). This is not recommended. The C++17 draft states that main() "shall not be used within a program." In other words, cannot be called from within a program. See e.g. Working Draft Standard for C++ Programming Language, dated "2017-03-21", Paragraph 6.6.1.3, page 66. I realize that some compilers support this (including mine), but the next version of the compiler could modify or remove that behavior as the standard uses the term "shall not".
It is illegal to call main from inside your program. That means the only thing that is going to call it is the runtime and the compiler/linker can handle setting that up.This means you do not need a prototype for main.
A definition of a function also implicitly declares it. If you need to reference a function before it is defined you need to declare it before you use it.
So writing the following is also valid:
int sum(int x, int y) {
return x + y;
}
int main() {
std::cout << "The result is " << sum(1, 2);
return 0;
}
If you use a declaration in one file to make a function known to the compiler before it is defined, then its definition has to be known at linking time:
main.cpp
int sum(int x, int y);
int main() {
std::cout << "The result is " << sum(1, 2);
return 0;
}
sum.cpp
int sum(int x, int y) {
return x + y;
}
Or sum could have its origin in a library, so you do not even compile it yourself.
The main function is not used/referenced in your code anywhere, so there is no need to add the declaration of main anywhere.
Before and after your main function the c++ library might execute some init and cleanup steps, and will call your main function. If that part of the library would be represented as c++ code then it would contain a declaration of int main() so that that it could be compiled. That code could look like this:
int main();
int __main() {
__startup_runtime();
main();
__cleanup_runtime();
}
But then you again have the same problem with __main so at some point there is no c++ anymore and a certain function (main) just represents the entry point of your code.
Nope. You can't call it anyway.
You only need forward declarations for functions called before they are defined. You need external declarations (which look exactly like forward declarations on purpose) for functions defined in other files.
But you can't call main in C++ so you don't need one. This is because the C++ compiler is allowed to modify main to do global initialization.
[I looked at crt0.c and it does have a declaration for main but that's neither here nor there].
Consider the following code:
file_1.hpp:
typedef void (*func_ptr)(void);
func_ptr file1_get_function(void);
file1.cpp:
// file_1.cpp
#include "file_1.hpp"
static void some_func(void)
{
do_stuff();
}
func_ptr file1_get_function(void)
{
return some_func;
}
file2.cpp
#include "file1.hpp"
void file2_func(void)
{
func_ptr function_pointer_to_file1 = file1_get_function();
function_pointer_to_file1();
}
While I believe the above example is technically possible - to call a function with internal linkage only via a function pointer, is it bad practice to do so? Could there be some funky compiler optimizations that take place (auto inline, for instance) that would make this situation problematic?
There's no problem, this is fine. In fact , IMHO, it is a good practice which lets your function be called without polluting the space of externally visible symbols.
It would also be appropriate to use this technique in the context of a function lookup table, e.g. a calculator which passes in a string representing an operator name, and expects back a function pointer to the function for doing that operation.
The compiler/linker isn't allowed to make optimizations which break correct code and this is correct code.
Historical note: back in C89, externally visible symbols had to be unique on the first 6 characters; this was relaxed in C99 and also commonly by compiler extension.
In order for this to work, you have to expose some portion of it as external and that's the clue most compilers will need.
Is there a chance that there's a broken compiler out there that will make mincemeat of this strange practice because they didn't foresee someone doing it? I can't answer that.
I can only think of false reasons to want to do this though: Finger print hiding, which fails because you have to expose it in the function pointer decl, unless you are planning to cast your way around things, in which case the question is "how badly is this going to hurt".
The other reason would be facading callbacks - you have some super-sensitive static local function in module m and you now want to expose the functionality in another module for callback purposes, but you want to audit that so you want a facade:
static void voodoo_function() {
}
fnptr get_voodoo_function(const char* file, int line) {
// you tagged the question as C++, so C++ io it is.
std::cout << "requested voodoo function from " << file << ":" << line << "\n";
return voodoo_function;
}
...
// question tagged as c++, so I'm using c++ syntax
auto* fn = get_voodoo_function(__FILE__, __LINE__);
but that's not really helping much, you really want a wrapper around execution of the function.
At the end of the day, there is a much simpler way to expose a function pointer. Provide an accessor function.
static void voodoo_function() {}
void do_voodoo_function() {
// provide external access to voodoo
voodoo_function();
}
Because here you provide the compiler with an optimization opportunity - when you link, if you specify whole program optimization, it can detect that this is a facade that it can eliminate, because you let it worry about function pointers.
But is there a really compelling reason not just to remove the static from infront of voodoo_function other than not exposing the internal name for it? And if so, why is the internal name so precious that you would go to these lengths to hide that?
static void ban_account_if_user_is_ugly() {
...;
}
fnptr do_that_thing() {
ban_account_if_user_is_ugly();
}
vs
void do_that_thing() { // ban account if user is ugly
...
}
--- EDIT ---
Conversion. Your function pointer is int(*)(int) but your static function is unsigned int(*)(unsigned int) and you don't want to have to cast it.
Again: Just providing a facade function would solve the problem, and it will transform into a function pointer later. Converting it to a function pointer by hand can only be a stumbling block for the compiler's whole program optimization.
But if you're casting, lets consider this:
// v1
fnptr get_fn_ptr() {
// brute force cast because otherwise it's 'hassle'
return (fnptr)(static_fn);
}
int facade_fn(int i) {
auto ui = static_cast<unsigned int>(i);
auto result = static_fn(ui);
return static_cast<int>(result);
}
Ok unsigned to signed, not a big deal. And then someone comes along and changes what fnptr needs to be to void(int, float);. One of the above becomes a weird runtime crash and one becomes a compile error.
There are a few legal ways which can we declare a function in C++.
Some of the legal ways are:
void function ();
void function (void);
dataType function (dataType);
and so on...
Recently, I came across a function declaration as such:
void (function) (); //Take note of the braces around the function name
I have never seen somehting like this before, when I tested it in a C++ compiler, it runs without any warning or compilation errors.
My question is: Why is void (function) (); a legal way to decalre a function prototype? Is there any special meaning to declare a function in this way? Or does it just work normally like any other function declaration?
One difference is that enclosing it in parenthesis prevents the function-like macros expansion by the preprocessor. As mentioned in the other answers it makes no difference though to the actual compiler.
For instance:
// somewhere buried deep in a header
#define function(a, b) a + b
// your code
void function() { // this expands the macro and gives compilation error
}
void (function)() { // this does not expand and works as expected
}
This comes in handy for instance when the bright minds behind the Microsoft Visual Studio library decided to provide function-like macros for things like min and max. (There are other ways like #undef to go around this).
Note that object-like macros (e.g. #define function 3 + 4) are still expanded.
The preprocessor is just a dumb text replacement tool (as opposed to the compiler which is just a (smart) text replacement tool). It takes the macro definition and replaces it everywhere. He is not aware of the semantics of what he replaces.
For instance:
// somewhere buried deep in a header
#define function 3 + 2
// your code
void function() {
}
The preprocessor sees the word function and textually replaces it with the string 3 + 2. He is unaware that function is a id-name part of a function declaration and definition. After the preprocess phase there come the actual compile phases. So the compiler actually sees:
// your code
void 3 + 2() {
}
which does not make any sense to him and gives an error.
For function-like macros
// somewhere buried deep in a header
#define function(a, b) a + b
The preprocessor does the same except that it expects two ‘tokens’ enclosed in parenthesis separated by comma (the parameters) and does the replacement. (again no semantics aware):
int d = function(2, 3);
//will be replaced by the preprocessor to:
int d = 2 + 3; // passes compilation phase
void function();
// the preprocessor doesn’t find the arguments for function so it gives an error.
However if it encounters (function) it will not try to expand it (it ignores it). It is just a rule.
it's the same as
void function();
you can declare it as
void ((function)) ();
if you want :)
be careful not to mix this up with the function pointer declaration syntax.
There is nothing special about it, it means exactly the same as the version without parentheses. It is just an artifact of how the syntax is declared. Usually you see the use of parentheses around the function name when a function pointer is declared, e.g.
void (*function_pointer)() = nullptr;
// a function pointer to a function taking and returning void
in contrast to
void *function();
// a function declaration of a function taking void and returning void*
I think it works the same as a normal function because function pointers are declared like: void (*function)() so if you leave out the * then it should be just a function.
It corresponds to the C++ grammar. If to simplify then one of the rules for defining of the declarator looks as
declarator:
(declarator)
So you can write for example
void (function) ();
or
void ( (function) () );
or even the following way
struct A
{
void ( ( function )() const );
};
I think you may find that was:
void (*function) ();
since there is no benefit to using void (function)(); or void (((((function)))))(); for that matter, they're equivalent. If I'm mistaken and it's not a typo, the answer is that you can put as many parentheses around the function name as you like, subject to compiler limitations, as per the code for output6() below.
If I'm not mistaken, that one with the * actually declares a function pointer which can be used to hold a pointer to a function. It does not declare a function at all, just a pointer that can be used to reference a function.
Like an int pointer (for example), the function pointer can point to an arbitrary function, parameters notwhithstanding.
So for example:
#include <iostream>
void (((((output6)))))() { std::cout << 6; }
void output7() { std::cout << 7; }
void output8() { std::cout << 8; }
void (*fn)();
int main() {
fn = &output6; fn();
fn = &output7; fn();
fn = &output8; fn();
std::cout << '\n';
}
would output 678.
I'm working on a project that delivers statistics to the user. I created a class called Dog,
And it has several functions. Speak, woof, run, fetch, etc.
I want to have a function that spits out how many times each function has been called. I'm also interested in the constructor calls and destructor calls as well.
I have a header file which defines all the functions, then a separate .cc file that implements them. My question is, is there a way to keep track of how many times each function is called?
I have a function called print that will fetch the "statistics" and then output them to standard output. I was considering using static integers as part of the class itself, declaring several integers to keep track of those things. I know the compiler will create a copy of the integer and initialize it to a minimum value, and then I'll increment the integers in the .cc functions.
I also thought about having static integers as a global variable in the .cc. Which way is easier? Or is there a better way to do this?
Any help is greatly appreciated!
Using static member variables is the way to go. However, the compiler will not "create a copy of the integer and initialize it to a minimum value"; you'll have to provide a definition for each one in the .cc file and initialize it to 0 there. (Things are a bit different if you're using C++11, but the basic idea is the same.)
There's no reason to use static global variables instead of static members.
foo.h:
class Foo {
static int countCtor_;
static int countDtor_;
static int countprint_:
Foo();
~Foo();
static void print();
};
foo.cc:
#include <iostream>
#include "foo.h"
int Foo::countCtor_ = 0;
int Foo::countDtor_ = 0;
int Foo::countprint_ = 0;
Foo::Foo() {
++countCtor_;
// Something here
}
Foo::~Foo() {
++countDtor_;
// Something here
}
void Foo::print() {
++countprint_;
std::cout << "Ctor: " << countCtor_ << "\n"
<< "Dtor: " << countDtor_ << "\n"
<< "print: " << countprint_ << "\n";
}
But if you've got a lot of functions, the repetition involved is a bit annoying—it's very easy to accidentally do ++countBar_ when you meant ++countBaz_ (especially if you copy and paste the boilerplate), so you may want something a bit fancier, such as a static map and a macro that increments counts[__FUNC__], so you can just use the exact same line in each function. Like this:
foo.h:
#include <map>
class Foo {
static std::map<const char*, int> counts_;
Foo();
~Foo();
void print();
};
foo.cc:
#include <iostream>
#include "foo.h"
std::map<const char *, int> Foo::counts_;
#define INC_COUNT_() do { ++counts_[__FUNC__]; } while (0)
Foo::Foo() {
INC_COUNT_();
// Something here
}
Foo::~Foo() {
INC_COUNT_();
// Something here
}
void Foo::print() {
INC_COUNT_();
for (std::map<const char *, int>::const_iterator it = counts_.begin();
it != counts_.end(); ++it) {
std::cout << it->first << ": " << it->second << "\n";
}
}
In the example code above, __FUNC__ is a placeholder. Unfortunately, there is no standard-compliant value you can use in its place. Most compilers have some subset of __func__, __FUNC__, __FUNCTION__, __FUNCSIG__, and __PRETTY_FUNCTION__. However, none of those are standard in C++03. C++11 does standardize __func__, but only as an "implementation-defined string", which isn't guaranteed to be useful, or even unique. On top of that, the values will be different on different compilers. Also, some of them may be macros rather than identifiers, to make things more fun.
If you want truly portable code, in C++11, you can use something like string(__func__) + ":" + STRINGIZE(__LINE__)—this will be somewhat ugly, but at least each function will have a unique name. And in C++03, there is no equivalent. If you just need "portable enough", consult the documentation for every compiler you use, or rely on something like autoconf.
Is there any reason you can't use standard profiling tools that will count these calls for you? Something like gprof?
Otherwise static integers would be the way to go.
Assuming you want these statistics tracked all the time in your program, you could use an unordered_map of your function names:
std::unordered_map<const char *, unsigned> stats;
void foo () {
// use __FUNCDNAME__ for MSVC
++stats[__PRETTY_FUNCTION__];
//...
}
The use of compiler specific function name specifiers is purposefully there to get the decorated function names. This is so that overloaded function names get counted as separate functions.
This technique allows you to add new functions easily without thinking about anything else, but there is a small additional cost if there are hash collisions (which can be remedied somewhat by sizing the stats map to be larger). There is no hash computed on the string, since the key is a pointer type, it just uses the pointer value itself as the hash.
If this is just one-off code for profiling, then you should first try to use the code profiling tools available on your platform.
You can put static locals inside the methods themselves, that seems cleaner since these variables aren't logically connected to the class so there's no reason to make them members.
Additionaly, you could have a macro to simplify the work. I normally don't recommend using macros, but this seems like an appropriate use:
#define DEFINE_COUNTER \
static int noCalls = 0; \
noCalls++;
void foo()
{
DEFINE_COUNTER
}
Use a library that implements the Observer Pattern or Method Call Interception. You can choose one from this list, or use something like Vitamin.
Here goes newbie question number 5, but I don't have a teacher.. so.. anyhow here we go:
I'm wondering if is necessary to have function prototypes at the top of the file, instead of putting the main function to the end of the file and create all the function at the top of the file. As far as I can tell, VC++ and G++ both are not complaining. Is there standards that disallows me to do so?
It seems rather annoying to have to change the prototype when you change a function parameters and return types.
Example:
#include <iostream>
void say_hi(){
std::cout << "hi" << std::endl;
}
int main(){
say_hi();
return 0;
}
This declares but does not define the function say_hi:
void say_hi();
This both declares and defines the function say_hi:
void say_hi(){
std::cout << "hi" << std::endl;
}
You can declare a function many times; you can only define it once.
A function must be declared in the file before you can call it. A function must be defined somewhere--in the same file before or after you call it or maybe even in a different file.
So, yes, this is perfectly fine.
You are correct; if you define all your functions above where they are called, you don't need function prototypes. The actual function definition serves the same purpose as a separate declaration.
This works when you have tiny functions. It works less well when they get long. Or when you have more than one file of code. As a matter of style, many teachers demand that even tiny applications be written with the structure that serves large applications well.