Chaining calls to temporaries in C++ - c++

I have a class that does a transformation on a string, like so
class transer{
transer * parent;
protected:
virtual string inner(const string & s) = 0;
public:
string trans(const string & s) {
if (parent)
return parent->trans(inner(s));
else
return inner(s);
}
transer(transer * p) : parent(p) {}
template <class T>
T create() { return T(this); }
template <class T, class A1> // no variadic templates for me
T create(A1 && a1) { return T(this, std::forward(a1)); }
};
So I can create a subclass
class add_count : public transer{
int count;
add_count& operator=(const add_count &);
protected:
virtual string inner(const string & s) {
return std::to_string((long long)count++) + s;
}
public:
add_count(transer * p = 0) : transer(p), count(0) {}
};
And then I can use the transformations:
void use_transformation(transer & t){
t.trans("string1");
t.trans("string2");
}
void use_transformation(transer && t){
use_trasnformation(t);
}
use_transformation(add_count().create<add_count>());
Is there a better design for this? I'd like to avoid using dynamic allocation/shared_ptr if I can, but I'm not sure if the temporaries will stay alive throughout the call. I also want to be able to have each transer be able to talk to its parent during destruction, so the temporaries also need to be destroyed in the right order. It's also difficult to create a chained transformation and save it for later, since
sometrans t = add_count().create<trans1>().create<trans2>().create<trans3>();
would save pointers to temporaries that no longer exist. Doing something like
trans1 t1;
trans2 t2(&t1);
trans3 t3(&t2);
would be safe, but annoying. Is there a better way to do these kinds of chained operations?

Temporaries will be destructed at the end of the full expression, in the
reverse order they were constructed. Be careful about the latter,
however, since there are no guarantees with regards to the order of
evaluation. (Except, of course, that of direct dependencies: if you
need one temporary in order to create the next—and if I've
understood correctly, that's your case—then you're safe.)

If you don't want dynamic allocation you either pass the data which is operated on to the function that initiates the chain, or you need a root type which holds it for you ( unless you want excessive copying ). Example ( might not compile ):
struct fooRef;
struct foo
{
fooRef create() { return fooRef( m_Val ); }
foo& operator=( const fooRef& a_Other );
std::string m_Val;
}
struct fooRef
{
fooRef( std::string& a_Val ) : m_Val( a_Val ) {}
fooRef create() { return fooRef( m_Val ); }
std::string& m_Val;
}
foo& foo::operator=( const fooRef& a_Other ) { m_Val = a_Other.m_Val; }
foo startChain()
{
return foo();
}
foo expr = startChain().create().create(); // etc
First the string lies on the temporary foo created from startChain(), all the chained operations operates on that source data. The assignment then at last copies the value over to the named var. You can probably almost guarantee RVO on startChain().

Related

How to make a hierarchy of different object "generators" in plain C++11

What I need is the following hierarchy of classes (given here as a sketch):
class DataClass {}
class AbstractGenerator {
public:
// Generates DataClass objects one by one. In a lazy manner
virtual DataClass produce() = 0;
}
class RandGenerator : AbstractGenerator {
public:
RandGenerator(int maximal_int) : maximal(maximal_int) {}
DataClass produce() {
// get a random number from 0 to this->maximal
// make a DataClass object from the random int and return it
}
private:
int maximal;
}
class FromFileGenerator : AbstractGenerator {
public:
FromFileGenerator(string file_name) : f_name(file_name) {}
DataClass produce() {
// read the next line from the file
// deserialize the DataClass object from the line and return it
}
private:
string f_name;
}
What I want to support for both RandGenerator and FromFileGnerator is:
RandGenerator rg();
for (DataClass data : rg) {...}
And also some method of taking "first n elements of the generator".
What are the appropriate tools in the plain C++11 that one could use to achieve this, or whatever is the closest to this in C++11?
boost::function_input_iterator is the normal tool for this job, but since you want "plain" C++, we can just reimplement the same concept.
class generator_iterator {
std::shared_ptr<AbstractGenerator> generator;
public:
using iterator_category = std::input_iterator_tag;
generator_iterator(std::shared_ptr<AbstractGenerator> generator_)
:generator(generator_) {}
DataClass operator*(){return generator->produce();}
generator_iterator& operator++(){return *this;}
generator_iterator operator++(int){return *this;}
//plus all the other normal bits for an output_iterator
};
And then in your AbstractGenerator class, provide begin and end methods.
generator_iterator begin() {return {this};}
generator_iterator end() {return {nullptr};} //or other logic depending on how you want to detect the end of a series
Add a begin and end member function to AbstractGenerator, which return iterators which call the produce member function.
Such an iterator (demo) could look similar to this:
template<typename Fn>
struct CallRepeatedlyIterator
{
using iterator_category = std::input_iterator_tag;
using value_type = typename std::result_of<Fn()>::type;
// Not sure if that's correct (probably not):
using difference_type = unsigned;
using pointer = value_type *;
using reference = value_type &;
bool is_end;
union {
Fn fn;
};
union {
value_type buffer;
};
value_type operator*() const
{
return buffer;
}
CallRepeatedlyIterator & operator++()
{
buffer = fn();
return *this;
}
CallRepeatedlyIterator()
: is_end(true)
{}
explicit CallRepeatedlyIterator(Fn f)
: is_end(false)
{
new (&fn) Fn(f);
new (&buffer) value_type(fn());
}
bool operator==(CallRepeatedlyIterator const & other) const
{
return is_end && other.is_end;
}
bool operator!=(CallRepeatedlyIterator const & other) const
{
return !(*this == other);
}
// NOTE: A destructor is missing here! It needs to destruct fn and buffer
// if its not the end iterator.
};
Now your begin member function returns such an iterator which calls produce (e.g. using a by reference capturing lambda) and end returns an "end" iterator.
This means that your for loop would run forever (no way to reach the end iterator)!

Proper API for access data members of container<T>

I have the following class:
class Document
{
public:
Document():
// default values for members,
// ...
m_dirty{false}{}
// Accessor functions
template<class OutputStream>
Document& save(OutputStream stream)
{
// Write stuff to `stream`
// ...
m_dirty = false;
return *this;
}
bool dirty() const { return m_dirty; }
private:
Size2d m_canvas_size;
LayerStack m_layers;
LayerIndex m_current_layer;
std::vector<Palette> m_palettes;
PaletteIndex m_current_palette;
ColorIndex m_current_color;
std::vector<std::string> m_palette_names;
std::vector<std::string> m_layer_names;
bool m_dirty;
};
Should the class have public member functions for modifying an element of say m_palettes directly, like
Document& color(PaletteIndex, ColorIndex, Color)
, or is it more "correct", to only allow access to the entire vector, through a pair of API:s
std::vector<Palette> const& palettes();
Document& palettes(std::vector<Palette>&&);
The first option would be more efficient, since it would not require to create a temporary copy of the data member, but consistent use of this design would make the interface bloated. It would require "deep" getters and setters for every container in the class.
Notice the dirty flag. Thus, the following would break the abstraction:
std::vector<Palette>& palettes();
You might have Proxy to "propagate" dirty flag from Palette modification, something like:
template <typename T>
class DirtyProxy
{
T& data;
bool& dirty;
public:
DirtyProxy(T& data, bool& dirty) : data(data), dirty(dirty) {}
~DirtyProxy() { dirty = true;}
DirtyProxy(const DirtyProxy&) = delete;
T* operator ->() { return data; }
};
And then
DirtyProxy<Palette> palette(std::size_t i) { return {m_palettes.at(i), dirty}; }
I think the most robust way to solve it is to use a a callback. An issue with the proxy is that it would not handle the case where the the client code throws an exception (assuming strong exception guarantee). Testcase:
try
{
auto property_proxy = obj.getProperty();
// an exception is thrown here...
property_proxy->val = x; // Never updated
}
catch(...)
{}
assert(!obj.dirty());
will fail, because the dtor always sets the dirty flag. However with a callback
class Foo
{
public:
template<class F>
Foo& modifyInSitu(F&& f)
{
f(x);
m_dirty = true;
return *this
}
};
will only update m_dirty, when f(x) does not throw.

Overload -> operator to forward member-access through Proxy

I'm trying to wrap a Python PyObject* in an Object class.
In Python, everything is a PyObject*.
A list is a PyObject*, and each item in the list is itself a PyObject*.
Which could even be another list.
etc.
I'm trying to allow fooList[42] = barObj style syntax by means of a Proxy pattern (here).
Now that I have that working, I want to extend it so that fooList[42] can be used as an Object. Specifically I want to be able to handle...
fooList[42].myObjMethod()
fooList[42].myObjMember = ...
Object has a lot of methods, and currently fooList[42].myObjMethod() is going to first resolve fooList[42] into a Proxy instance, say tmpProxy, and then attempt tmpProxy.myObjMethod().
This means I would have to do
void Proxy::myObjMethod(){ return wrapped_ob.myObjMethod(); }
i.e. manually relay each of Object's methods through Proxy, which is ugly.
I can't see any perfect solution (see the above linked answer), but I would be happy to use:
fooList[42]->myObjMethod()
... as a compromise, seeing as -> can be overloaded (as opposed to . which cannot).
However, I can't find any documentation for overloading operator->.
My best guess is that it must return a pointer to some object (say pObj), and C++ will invoke pObj->whatever.
Below is my attempted implementation. However, I'm running into a 'taking the address of a temporary object of type Object' warning.
I have, within my Object class:
const Object operator[] (const Object& key) const {
return Object{ PyObject_GetItem( p, key.p ) };
}
NOTE that 'const Object&' runs into 'taking the address of a temporary object of type Object' warning.
class Proxy {
private:
const Object& container;
const Object& key;
public:
// at this moment we don't know whether it is 'c[k] = x' or 'x = c[k]'
Proxy( const Object& c, const Object& k ) : container{c}, key{k}
{ }
// Rvalue
// e.g. cout << myList[5] hits 'const Object operator[]'
operator Object() const {
return container[key];
}
// Lvalue
// e.g. (something = ) myList[5] = foo
const Proxy& operator= (const Object& rhs_ob) {
PyObject_SetItem( container.p, key.p, rhs_ob.p );
return *this; // allow daisy-chaining a = b = c etc, that's why we return const Object&
}
const Object* operator->() const { return &container[key]; }
// ^ ERROR: taking the address of a temporary object of type Object
};
The idea is to allow myList[5]->someMemberObj = ... style syntax.
myList[5] resolves as a Proxy instance, which is wrapping an Object (the sixth element of myList). Let's call it myItem.
Now I want someProxy->fooFunc() or someProxy->fooProperty to invoke myItem.fooFunc() or myItem.fooProperty respectively.
I'm running into a 'taking the address of a temporary object of type Object' warning.
If you can change Object, you may add
class Object {
public:
// other code
const Object* operator -> () const { return this; }
Object* operator -> () { return this; }
};
And for your Proxy
Object operator->() { return container[key]; }
So, for example
myObj[42]->myFoo = ...
is mostly equivalent to
Proxy proxy = myObj[42];
Object obj = proxy.operator ->();
Object* pobj = obj.operator ->(); // so pobj = &obj;
pobj->myFoo = ...
I find the Proxy class that you wrote as an example a bit confusing so i took the liberty to change it a little:
Here is a simple example:
//object with lots of members:
class my_obj
{
public:
std::string m_text;
void foo()
{
std::cout << m_text << std::endl;
}
void bar(std::string t)
{
m_text = t;
}
};
//proxy object
class proxy_class
{
private:
friend class CustomContainer;
my_obj* px;
proxy_class(my_obj * obj_px)
:px(obj_px)
{
}
proxy_class() = delete;
proxy_class(const proxy_class &) = delete;
proxy_class& operator =(const proxy_class &) = delete;
public:
my_obj* operator ->()
{
return px;
}
};
//custom container that is the only one that can return proxy objects
class CustomContainer
{
public:
std::map<std::size_t, my_obj> stuff;
proxy_class operator [](const std::size_t index)
{
return proxy_class(&stuff[index]);
}
};
example usage:
CustomContainer cc;
cc[0]->foo();
cc[0]->bar("hello world");
cc[0]->foo();
As a design consideration the proxy class should be create in a controlled environment so constructors are removed from preventing miss-usage.
CustomContainer has to only return proxy_class with a reference to my_obj so it can use anything, std::map, std::vector, etc
After several hours of coaxing coliru, I have a working testcase.
Please refer to: https://codereview.stackexchange.com/questions/75237/c11-proxy-pattern-for-supporting-obidx-someobjmember-type-acc
Many thanks to Jarod, for supplying the correct syntax and understanding for -> overload.

Is there an idiomatic way to return a pointer that optionally owns its value

I have a function that given a path name, does a look up and returns a pointer to the associated value. Sometimes the value lives in a static cache, sometimes it gets calculated and created on the fly.
So, sometimes the caller takes ownership and needs to delete the object after reading it, and sometimes not. I'm wondering, is there something I can wrap this pointer with so that it will automatically be freed as necessary by the caller?
I was thinking I might be able to use a unique_ptr, but isn't the deleter part of the type, so how could I return the same type that sometimes does and sometimes doesn't actually delete.
So indeed, one solution could be returning a normal std::shared_ptr for the value created inside the function, and another one with an empty deleter for the value that lives in the map.
Live example of this solution
You can see how both use cases don't require any actions from the calling code and are completely transparent.
You can use std::unique_ptr with a deleter that knows whether to free or not. While the deleter type is part of the unique_ptr type, different unique_ptr instances can have different deleter instances:
template <class T>
class delete_if_not_cached {
bool cached;
public:
delete_if_not_cached(bool c = false) : cached(c) {}
void operator()(T *obj) { if (!cached) delete obj; }
}
and you have your function return a std::unique_ptr<T, delete_if_not_cached<T>>. If you're returning a pointer into the cache, you create that pointer as:
return std::unique_ptr<T, delete_if_not_cached<T>>(raw_pointer, delete_if_not_cached<T>(true));
to return a non-cached object, use
return std::unique_ptr<T, delete_if_not_cached<T>>(new T(...))
One potential pitfall is that if you ever remove things from the cache, that might leave dangling unique_ptrs that you have previously returned. If that's an issue, it probably makes more sense to use shared_ptrs both to return and in the cache itself.
You could use a std::shared_ptr but that does not really describe your ownership model. Have you considered rolling your own wrapper that contains a std::unique_ptr and a raw pointer and uses the correct one depending on the circumstances? Something like:
#include <cassert>
#include <memory>
class MyClass { };
class Wrapper {
const MyClass* cached_;
std::unique_ptr<MyClass> owned_;
public:
Wrapper() : cached_(nullptr) {}
void setCached(const MyClass* cached) {cached_ = cached;}
void setOwned(std::unique_ptr<MyClass> owned) { owned_ = std::move(owned); }
const MyClass* get() const {return cached_ ? cached_ : owned_.get();}
};
Wrapper getWrapper(int i) {
static MyClass first;
static MyClass second;
Wrapper wrapper;
if (i == 0)
wrapper.setCached(&first);
else if (i == 1)
wrapper.setCached(&second);
else
wrapper.setOwned(std::unique_ptr<MyClass>(new MyClass()));
return wrapper;
}
int main() {
for (int i = 0; i != 4; ++i) {
Wrapper wrapper = getWrapper(i);
assert(wrapper.get() != nullptr);
}
}
The wrapper can either forward calls to the real class or provide access to a raw pointer to the real class.
Or the wrapper could work polymorphically, with an interface and two implementations. One with a raw pointer and one with a unique pointer:
#include <cassert>
#include <memory>
class MyClass {};
class Wrapper {
public:
virtual ~Wrapper() = 0;
virtual const MyClass* get() const = 0;
};
Wrapper::~Wrapper() {}
class OwnerWrapper : public Wrapper {
std::unique_ptr<MyClass> owned_;
public:
OwnerWrapper(std::unique_ptr<MyClass> in) : owned_(std::move(in)) {}
virtual const MyClass* get() const { return owned_.get(); }
};
class PtrWrapper : public Wrapper {
const MyClass* ptr_;
public:
PtrWrapper(const MyClass* ptr) : ptr_(ptr) {}
virtual const MyClass* get() const { return ptr_; }
};
std::unique_ptr<Wrapper> getWrapper(int i) {
static MyClass first;
static MyClass second;
if (i == 0)
return std::unique_ptr<Wrapper>(new PtrWrapper(&first));
else if (i == 1)
return std::unique_ptr<Wrapper>(new PtrWrapper(&second));
else {
std::unique_ptr<MyClass> myclass(new MyClass());
return std::unique_ptr<Wrapper>(new OwnerWrapper(std::move(myclass)));
}
}
int main() {
for (int i = 0; i != 4; ++i) {
auto wrapper = getWrapper(i);
assert(wrapper->get() != nullptr);
}
}

What is purpose of a "this" pointer in C++? [duplicate]

This question already has answers here:
When should I make explicit use of the `this` pointer?
(12 answers)
Closed 6 years ago.
What is purpose of this keyword. Doesn't the methods in a class have access to other peer members in the same class ? What is the need to call a this to call peer methods inside a class?
Two main uses:
To pass *this or this as a parameter to other, non-class methods.
void do_something_to_a_foo(Foo *foo_instance);
void Foo::DoSomething()
{
do_something_to_a_foo(this);
}
To allow you to remove ambiguities between member variables and function parameters. This is common in constructors.
MessageBox::MessageBox(const string& message)
{
this->message = message;
}
(Although an initialization list is usually preferable to assignment in this particular example.)
Helps in disambiguating variables.
Pass yourself as a parameter or return yourself as a result
Example:
struct A
{
void test(int x)
{
this->x = x; // Disambiguate. Show shadowed variable.
}
A& operator=(A const& copy)
{
x = copy.x;
return *this; // return a reference to self
}
bool operator==(A const& rhs) const
{
return isEqual(*this, rhs); // Pass yourself as parameter.
// Bad example but you can see what I mean.
}
private:
int x;
};
Consider the case when a parameter has the same name as a class member:
void setData(int data){
this->data = data;
}
Resolve ambgiguity between member variables/functions and those defined at other scopes
Make explicit to a reader of the code that a member function is being called or a member variable is being referenced.
Trigger IntelliSense in the IDE (though that may just be me).
The expression *this is commonly used to return the current object from a member function:
return *this;
The this pointer is also used to guard against self-reference:
if (&Object != this) {
// do not execute in cases of self-reference
It lets you pass the current object to another function:
class Foo;
void FooHandler(Foo *foo);
class Foo
{
HandleThis()
{
FooHandler(this);
}
};
Some points to be kept in mind
This pointer stores the address of
the class instance, to enable pointer
access of the members to the member
functions of the class.
This pointer is not counted for
calculating the size of the object.
This pointers are not accessible for
static member functions.
This pointers are not modifiable
Look at the following example to understand how to use the 'this' pointer explained in this C++ Tutorial.
class this_pointer_example // class for explaining C++ tutorial
{
int data1;
public:
//Function using this pointer for C++ Tutorial
int getdata()
{
return this->data1;
}
//Function without using this pointer
void setdata(int newval)
{
data1 = newval;
}
};
Thus, a member function can gain the access of data member by either using this pointer or not.
Also read this to understand some other basic things about this pointer
It allows you to get around members being shadowed by method arguments or local variables.
The this pointer inside a class is a reference to itself. It's needed for example in this case:
class YourClass
{
private:
int number;
public:
YourClass(int number)
{
this->number = number;
}
}
(while this would have been better done with an initialization list, this serves for demonstration)
In this case you have 2 variables with the same name
The class private "number"
And constructor parameter "number"
Using this->number, you let the compiler know you're assigning to the class-private variable.
For example if you write an operator=() you must check for self assignment.
class C {
public:
const C& operator=(const C& rhs)
{
if(this==&rhs) // <-- check for self assignment before anything
return *this;
// algorithm of assignment here
return *this; // <- return a reference to yourself
}
};
The this pointer is a way to access the current instance of particular object. It can be used for several purposes:
as instance identity representation (for example in comparison to other instances)
for data members vs. local variables disambiguation
to pass the current instance to external objects
to cast the current instance to different type
One more purpose is to chaining object:
Consider the following class:
class Calc{
private:
int m_value;
public:
Calc() { m_value = 0; }
void add(int value) { m_value += value; }
void sub(int value) { m_value -= value; }
void mult(int value) { m_value *= value; }
int getValue() { return m_value; }
};
If you wanted to add 5, subtract 3, and multiply by 4, you’d have to do this:
#include
int main()
{
Calc calc;
calc.add(5); // returns void
calc.sub(3); // returns void
calc.mult(4); // returns void
std::cout << calc.getValue() << '\n';
return 0;
}
However, if we make each function return *this, we can chain the calls together. Here is the new version of Calc with “chainable” functions:
class Calc
{
private:
int m_value;
public:
Calc() { m_value = 0; }
Calc& add(int value) { m_value += value; return *this; }
Calc& sub(int value) { m_value -= value; return *this; }
Calc& mult(int value) { m_value *= value; return *this; }
int getValue() { return m_value; }
};
Note that add(), sub() and mult() are now returning *this. Consequently, this allows us to do the following:
#include <iostream>
int main()
{
Calc calc;
calc.add(5).sub(3).mult(4);
std::cout << calc.getValue() << '\n';
return 0;
}
We have effectively condensed three lines into one expression.
Copied from :http://www.learncpp.com/cpp-tutorial/8-8-the-hidden-this-pointer/
Sometimes you want to directly have a reference to the current object, in order to pass it along to other methods or to store it for later use.
In addition, method calls always take place against an object. When you call a method within another method in the current object, is is equivalent to writing this->methodName()
You can also use this to access a member rather than a variable or argument name that "hides" it, but it is (IMHO) bad practice to hide a name. For instance:
void C::setX(int x)
{
this->x = x;
}
For clarity, or to resolve ambiguity when a local variable or parameter has the same name as a member variable.
It also allows you to test for self assignment in assignment operator overloads:
Object & operator=(const Object & rhs) {
if (&rhs != this) {
// do assignment
}
return *this;
}
It also allows objects to delete themselves. This is used in smart pointers implementation, COM programming and (I think) XPCOM.
The code looks like this (excerpt from some larger code):
class counted_ptr
{
private:
counted_ptr(const counted_ptr&);
void operator =(const counted_ptr&);
raw_ptr_type _ptr;
volatile unsigned int _refcount;
delete_function _deleter;
public:
counted_ptr(raw_ptr_type const ptr, delete_function deleter)
: _ptr(ptr), _refcount(1), _deleter(deleter) {}
~counted_ptr() { (*_deleter)(_ptr); }
unsigned int addref() { return ++_refcount; }
unsigned int release()
{
unsigned int retval = --_refcount;
if(0 == retval)
>>>>>>>> delete this;
return retval;
}
raw_ptr_type get() { return _ptr; }
};
The double colon in c++ is technically known as "Unary Scope resolution operator".
Basically it is used when we have the same variable repeated for example inside our "main" function (where our variable will be called local variable) and outside main (where the variable is called a global variable).
C++ will alwaysexecute the inner variable ( that is the local one).
So imagine you want to use the global variable "Conundrum" instead the local one just because the global one is expressed as a float instead of as an integer:
#include <iostream>
using namespace std;
float Conundrum=.75;
int main()
{
int Conundrum =75;
cout<<::Conundrum;
}
So in this case the program will use our float Conundrum instead of the int Conundrum.