casting a returned value from C++ function - c++

I have following C++ code snippet
inline std::vector<std::unique_ptr<xir::Tensor>> cloneTensorBuffer(
const std::vector<const xir::Tensor*>& tensors)
{
auto ret = std::vector<std::unique_ptr<xir::Tensor>>{};
auto type = xir::DataType::XINT;
ret.reserve(tensors.size());
for (const auto& tensor : tensors) {
ret.push_back(std::unique_ptr<xir::Tensor>(xir::Tensor::create(
tensor->get_name(), tensor->get_shape(), xir::DataType{type, 8u})));
}
return ret;
}
I am not clear about the expression:
std::unique_ptr<xir::Tensor>(xir::Tensor::create(
tensor->get_name(), tensor->get_shape(), xir::DataType{type, 8u}))
Is the expression casting the value returned by xir::Tensor::create() to std::unique_ptr<xir::Tensor? I am confused since the C++ casting syntax is (type)expression
Can someone explain please.
regards,
-sunil puranik

This code takes a vector of raw pointers to Tensor objects, and returns a new vector of unique pointers to a new set of Tensor objects. Doesn't look like any explicit casting is taking place.
std::unique_ptr<xir::Tensor>( // create unique pointer to new Tensor object
xir::Tensor::create(
tensor->get_name(), // created with name and shape of the
tensor->get_shape(), // tensor pointed at in the original vector
xir::DataType{type, 8u} // and with unsigned 8-bit XINT datatype
)
)
Regarding casting - please don't consider e.g. (int)foo to be C++-style casting - that's legacy C-style casting. Amongst other things it's hard to find by searching a codebase!
C++ has:
static_cast
dynamic_cast
reinterpret_cast
const_cast
implicit conversion
There's more detail on this at cplusplus.com.

Related

Why can't we use a void* to operate on the object it addresses

I am learning C++ using C++ Primer 5th edition. In particular, i read about void*. There it is written that:
We cannot use a void* to operate on the object it addresses—we don’t know that object’s type, and the type determines what operations we can perform on that object.
void*: Pointer type that can point to any nonconst type. Such pointers may not
be dereferenced.
My question is that if we're not allowed to use a void* to operate on the object it addressess then why do we need a void*. Also, i am not sure if the above quoted statement from C++ Primer is technically correct because i am not able to understand what it is conveying. Maybe some examples can help me understand what the author meant when he said that "we cannot use a void* to operate on the object it addresses". So can someone please provide some example to clarify what the author meant and whether he is correct or incorrect in saying the above statement.
My question is that if we're not allowed to use a void* to operate on the object it addressess then why do we need a void*
It's indeed quite rare to need void* in C++. It's more common in C.
But where it's useful is type-erasure. For example, try to store an object of any type in a variable, determining the type at runtime. You'll find that hiding the type becomes essential to achieve that task.
What you may be missing is that it is possible to convert the void* back to the typed pointer afterwards (or in special cases, you can reinterpret as another pointer type), which allows you to operate on the object.
Maybe some examples can help me understand what the author meant when he said that "we cannot use a void* to operate on the object it addresses"
Example:
int i;
int* int_ptr = &i;
void* void_ptr = &i;
*int_ptr = 42; // OK
*void_ptr = 42; // ill-formed
As the example demonstrates, we cannot modify the pointed int object through the pointer to void.
so since a void* has no size(as written in the answer by PMF)
Their answer is misleading or you've misunderstood. The pointer has a size. But since there is no information about the type of the pointed object, the size of the pointed object is unknown. In a way, that's part of why it can point to an object of any size.
so how can a int* on the right hand side be implicitly converted to a void*
All pointers to objects can implicitly be converted to void* because the language rules say so.
Yes, the author is right.
A pointer of type void* cannot be dereferenced, because it has no size1. The compiler would not know how much data he needs to get from that address if you try to access it:
void* myData = std::malloc(1000); // Allocate some memory (note that the return type of malloc() is void*)
int value = *myData; // Error, can't dereference
int field = myData->myField; // Error, a void pointer obviously has no fields
The first example fails because the compiler doesn't know how much data to get. We need to tell it the size of the data to get:
int value = *(int*)myData; // Now fine, we have casted the pointer to int*
int value = *(char*)myData; // Fine too, but NOT the same as above!
or, to be more in the C++-world:
int value = *static_cast<int*>(myData);
int value = *static_cast<char*>(myData);
The two examples return a different result, because the first gets an integer (32 bit on most systems) from the target address, while the second only gets a single byte and then moves that to a larger variable.
The reason why the use of void* is sometimes still useful is when the type of data doesn't matter much, like when just copying stuff around. Methods such as memset or memcpy take void* parameters, since they don't care about the actual structure of the data (but they need to be given the size explicitly). When working in C++ (as opposed to C) you'll not use these very often, though.
1 "No size" applies to the size of the destination object, not the size of the variable containing the pointer. sizeof(void*) is perfectly valid and returns, the size of a pointer variable. This is always equal to any other pointer size, so sizeof(void*)==sizeof(int*)==sizeof(MyClass*) is always true (for 99% of today's compilers at least). The type of the pointer however defines the size of the element it points to. And that is required for the compiler so he knows how much data he needs to get, or, when used with + or -, how much to add or subtract to get the address of the next or previous elements.
void * is basically a catch-all type. Any pointer type can be implicitly cast to void * without getting any errors. As such, it is mostly used in low level data manipulations, where all that matters is the data that some memory block contains, rather than what the data represents. On the flip side, when you have a void * pointer, it is impossible to determine directly which type it was originally. That's why you can't operate on the object it addresses.
if we try something like
typedef struct foo {
int key;
int value;
} t_foo;
void try_fill_with_zero(void *destination) {
destination->key = 0;
destination->value = 0;
}
int main() {
t_foo *foo_instance = malloc(sizeof(t_foo));
try_fill_with_zero(foo_instance, sizeof(t_foo));
}
we will get a compilation error because it is impossible to determine what type void *destination was, as soon as the address gets into try_fill_with_zero. That's an example of being unable to "use a void* to operate on the object it addresses"
Typically you will see something like this:
typedef struct foo {
int key;
int value;
} t_foo;
void init_with_zero(void *destination, size_t bytes) {
unsigned char *to_fill = (unsigned char *)destination;
for (int i = 0; i < bytes; i++) {
to_fill[i] = 0;
}
}
int main() {
t_foo *foo_instance = malloc(sizeof(t_foo));
int test_int;
init_with_zero(foo_instance, sizeof(t_foo));
init_with_zero(&test_int, sizeof(int));
}
Here we can operate on the memory that we pass to init_with_zero represented as bytes.
You can think of void * as representing missing knowledge about the associated type of the data at this address. You may still cast it to something else and then dereference it, if you know what is behind it. Example:
int n = 5;
void * p = (void *) &n;
At this point, p we have lost the type information for p and thus, the compiler does not know what to do with it. But if you know this p is an address to an integer, then you can use that information:
int * q = (int *) p;
int m = *q;
And m will be equal to n.
void is not a type like any other. There is no object of type void. Hence, there exists no way of operating on such pointers.
This is one of my favourite kind of questions because at first I was also so confused about void pointers.
Like the rest of the Answers above void * refers to a generic type of data.
Being a void pointer you must understand that it only holds the address of some kind of data or object.
No other information about the object itself, at first you are asking yourself why do you even need this if it's only able to hold an address. That's because you can still cast your pointer to a more specific kind of data, and that's the real power.
Making generic functions that works with all kind of data.
And to be more clear let's say you want to implement generic sorting algorithm.
The sorting algorithm has basically 2 steps:
The algorithm itself.
The comparation between the objects.
Here we will also talk about pointer functions.
Let's take for example qsort built in function
void qsort(void *base, size_t nitems, size_t size, int (*compar)(const void *, const void*))
We see that it takes the next parameters:
base − This is the pointer to the first element of the array to be sorted.
nitems − This is the number of elements in the array pointed by base.
size − This is the size in bytes of each element in the array.
compar − This is the function that compares two elements.
And based on the article that I referenced above we can do something like this:
int values[] = { 88, 56, 100, 2, 25 };
int cmpfunc (const void * a, const void * b) {
return ( *(int*)a - *(int*)b );
}
int main () {
int n;
printf("Before sorting the list is: \n");
for( n = 0 ; n < 5; n++ ) {
printf("%d ", values[n]);
}
qsort(values, 5, sizeof(int), cmpfunc);
printf("\nAfter sorting the list is: \n");
for( n = 0 ; n < 5; n++ ) {
printf("%d ", values[n]);
}
return(0);
}
Where you can define your own custom compare function that can match any kind of data, there can be even a more complex data structure like a class instance of some kind of object you just define. Let's say a Person class, that has a field age and you want to sort all Persons by age.
And that's one example where you can use void * , you can abstract this and create other use cases based on this example.
It is true that is a C example, but I think, being something that appeared in C can make more sense of the real usage of void *. If you can understand what you can do with void * you are good to go.
For C++ you can also check templates, templates can let you achieve a generic type for your functions / objects.

Iterating through a vector of stucts's members with pointers and offsets Part 2

So to this is part 2 of a question I asked and was answered yesterday. So today I am coming back with a part 2. I am not sure if this should be this somewhere else so if a moderator wants to move it feel free.
So I am not going to reintroduce my problem here, so please go read part 1
Iterating through a vector of stucts's members with pointers and offsets
So I have come up with a solution to the problem so let me post a modified snippet of code that represents the solution I am going for,
#include <iostream>
#include <vector>
// knows nothing about foo
class Readfoo
{
private:
int offSetSize;
char* pchar;
public:
void SetPoint(double* apstartDouble, int aoffSetSize)
{
offSetSize = aoffSetSize;
pchar = static_cast<char*> (static_cast<void*>(apstartDouble));
};
const double& printfoo(int aioffset) const
{
return *(static_cast<double*> (static_cast<void*>(pchar + aioffset*offSetSize)));
};
};
// knows nothing about readFoo
struct foo
{
int a[5];
double b[10];
};
int main()
{
// populate some data (choose b [2] or other random entry.).
std::vector<foo> bar(10);
for(int ii = 0; ii < bar.size(); ii++)
bar[ii].b[2] = ii;
// access b[2] for each foo using an offset.
Readfoo newReadfoo;
newReadfoo.SetPoint(&(bar[0].b[2]), sizeof(foo)/sizeof(char));
for(int ii = 0; ii < bar.size(); ii++)
std::cout<<"\n"<<newReadfoo.printfoo(ii);
return 0;
}
This, in my opinion, is legal, well I suppose that is what I am asking. Since now, in essence, I am converting my 'interpretation' of the struct foo and the vector bar (an array of foos) into a single array of bytes, or chars.
I.e. In this interpretation the data structure is a single array of chars, of foo size times bar size. When I iterate through this with an integral type I am in essence moving to some hypothetical char element (point 4.2 in the answer to part 1). The printfoo function then combines the next 8 bytes to form a double to return.
So is this legal and other than moving out of bounds of the bar vectors is there any reason why this will not work (I have tested it and it has yet to fail.)?
So is this legal ...
No it is not.
In the following expression:
pchar + aioffset*offSetSize
you manipulate pchar as if it were a pointer to a char array, which it is not (it's a pointer to a double array). And this is undefined behavior:
[expr.add]/6
For addition or subtraction, if the expressions P or Q have type “pointer to cv T”, where T and the array element type are not similar, the behavior is undefined. [ Note: In particular, a pointer to a base class cannot be used for pointer arithmetic when the array contains objects of a derived class type. — end note ]
In your case P is pchar and has type pointer to char but the array element it points to is a double.
... and other than moving out of bounds of the bar vectors is there any reason why this will not work (I have tested it and it has yet to fail.)?
Yes: Does the C++ standard allow for an uninitialized bool to crash a program?
To go further: pointer manipulation in C++ is a red flag. C++ pointer manipulation is dark magic which will burn your soul and consume your dog. C++ offers a lot of tools to write generic code. My advice: Ask about what you're trying to achieve rather than about your attempted solution. You'll learn a lot.

invalid conversion from 'void* - casting

I have an issue with a code that compiles without problem under gcc but not with g++. I am moving to g++ since I would like to make use of boost for some multithreading improvement. The errors I get are something like:
invalid conversion from 'void* volatile' to 'TrialVect*'
the piece of code is:
static void funcSize( struct TrialVect *vector ) {
if ( vector->field ) {
struct TrialVect *ptr = BO2V (vector->first,vector->second);
}
}
As I googled there is some issue with casting variables, why? suggestions how to fix it?
In C++ any pointer can be converted to void* implicitly, but converting from void* requires an explicit cast.
Try this:
auto ptr = static_cast<TrialVect* volatile>( BO2V (vector->first,vector->second) );
volatile is an attribute (like const) that must be handled separately; either the new pointer must match the old one, or you can change the const/volatile attributes using a separate const_cast. For example,
auto ptr = static_cast<TrialVect*>( const_cast<void*>( BO2V(vector->first, vector->second) ) );
Whether removing the volatile attribute is safe or not depends on the application.

Convert void* to a dynamic type

I have a double variable i, that is converted to a void pointer named pointer:
double i = 3;
void *pointer = &i;
When I like to convert the void pointer back to double I used:
double p = *((double*) pointer);
I would like to convert the void pointer to the type I will send as a char*:
char* myType= typeid(i).name();//get the name of type of "i"
int p = *((myType*) pointer); // how to implement?
Is it possible?
instead of
char* myType= typeid(i).name();//get the name of type of "i"
int p = *((myType*) pointer); // how to implement?
use
typedef decltype(i) myType;
myType p = *((myType*) pointer);
or better:
typedef decltype(i) myType;
auto p = *reinterpret_cast<myType*>(pointer);
Works with c++11 or later. If you want to decltype on older c++ compilers, it is emulated in boost.
Edit. This is probably different from what you wanted to do, which I suppose is something like this:
void myFunction(void* unknownParam) {
typedef (realTypeOf unknownParam) RealType; // <-- this is not real c++
RealType &a = *reinterpret_cast<RealType*>(unknownParam)
//do stuff using 'a'
}
This is not possible in C++, but there is a reason: it doesn't make much sense.
And the reason is that for myFunction to be valid the //do stuff using 'a' part should be valid for whatever type RealType ends up being. As such, it cannot rely on any feature the RealType type have: it cannot use any of its methods, it cannot use any operator, it cannot even know whether it is a class or not. Basically, you cannot do anything more with it than what you would already be able to do on a void*, so giving the type a name doesn't really help you much.
A language feature that is similar to what you want (but not quite it) is type reflection, which is not present in C++, but you can find it in language such as Java, Objective-C or C#. Basically, you ask the object itself if it has a certain method, and eventually call it. An example in Objective-C
-(void)myFunction:(id)unknownParam {
if([unknownParam respondsToSelector:#selector(myMethod)])
[unknownParam performSelector:#selector(myMethod)]
}
C/C++ does not work well to interchange datatype like for example JavaScript variables
Format of int value will be different than double value format (floating point) in binary
You cannot get original data type using typeid after it has been casted to void*. Also note that typeid will have different output on different OS and compilers
double dValue = 77.7;
void* pValue = &dValue;
//output "pV" / pointer void (depending on compiler and OS)
std::cout << typeid(dValue).name() << std::endl;
To cast from void* using string you can make rules like following. Or you can try to use C++ template functions in specific cases.
int iOutValue = 0;
double dOutValue = 0;
char* name = "double";
if(!strcmp(name, "int"))
{
iOutValue = *((int*)pValue);
}
else if(!strcmp(name, "double"))
{
dOutValue = *((double*)pValue);
}
If instead of passing around void* you used some kind of variant type, you would be able to convert it back.
If you use a string to indicate the type you will need some kind of map from string to actual type. Although you can go from type to string, there is no conversion back.
Assuming you know your underlying data is always numeric in some way, there are ways to have special discrete variants that only contain numbers. The simplest would be to store a union of a 64-bit int and a double, and some flag indicating which one you have, and then a method to convert to any numeric type, asDouble(), asLong() etc.

C++ static_cast from int* to void* to char* - can you help me to understand this code?

I'm a beginner in C++, and I have problem with understanding some code.
I had an exercise to do, to write function which returns size of int, and do not use sizeof() and reinterpret_cast. Someone gave me solution, but I do not understand how it works. Can you please help me to understand it? This is the code:
int intSize() {
int intArray[10];
int * intPtr1;
int * intPtr2;
intPtr1 = &intArray[1];
intPtr2 = &intArray[2];
//Why cast int pointer to void pointer?
void* voidPtr1 = static_cast<void*>(intPtr1);
//why cast void pointer to char pointer?
char* charPtr1 = static_cast<char*>(voidPtr1);
void* voidPtr2 = static_cast<void*>(intPtr2);
char* charPtr2 = static_cast<char*>(voidPtr2);
//when I try to print 'charPtr1' there is nothing printed
//when try to print charPtr2 - charPtr1, there is correct value shown - 4, why?
return charPtr2 - charPtr1;
}
To summarize what I don't understand is, why we have to change int* to void* and then to char* to do this task? And why we have the result when we subtract charPtr2 and charPtr1, but there is nothing shown when try to print only charPtr1?
First of all, never do this in real-world code. You will blow off your leg, look like an idiot and all the cool kids will laugh at you.
That being said, here's how it works:
The basic idea is that the size of an int is equal to the offset between two elements in an int array in bytes. Ints in an array are tightly packed, so the beginning of the second int comes right after the end of the first one:
int* intPtr1 = &intArray[0];
int* intPtr2 = &intArray[1];
The problem here is that when subtracting two int pointers, you won't get the difference in bytes, but the difference in ints. So intPtr2 - intPtr1 is 1, because they are 1 int apart.
But we are in C++, so we can cast pointers to anything! So instead of using int pointers, we copy the value to char pointers, which are 1 byte in size (at least on most platforms).
char* charPtr1 = reinterpret_cast<char*>(intPtr1);
char* charPtr2 = reinterpret_cast<char*>(intPtr2);
The difference charPtr2 - charPtr1 is the size in bytes. The pointers still point to the same location as before (i.e. the start of the second and first int in the array), but the difference will now be calculated in sizes of char, not in sizes of int.
Since the exercise did not allow reinterpret_cast you will have to resort to another trick. You cannot static_cast from int* to char* directly. This is C++'s way of protecting you from doing something stupid. The trick is to cast to void* first. You can static_cast any pointer type to void* and from void* to any pointer type.
This is the important bit:
intPtr1 = &intArray[1];
intPtr2 = &intArray[2];
This creates two pointers to adjacent ints in the array. The distance between these two pointers is the size of an integer that you're trying to retrieve. However the way that pointer arithmetic works is that if you subtract these two then the compiler will return you the size in terms of ints, which will always be 1.
So what you're doing next is re-casting these as character pointers. Characters are (or de-facto are) 1 byte each, so the difference between these two pointers as character pointers will give you an answer in bytes. That's why you're casting to character pointers and subtracting.
As for via void* - this is to avoid having to use reinterpret_cast. You're not allowed to cast directly from a int* to a char* with static_cast<>, but going via void* removes this restriction since the compiler no longer knows it started with an int*. You could also just use a C-style cast instead, (char*)(intPtr1).
"do not use sizeof() and reinterpret_cast"... nothing's said about std::numeric_limits, so you could do it like that :)
#include <limits>
int intSize()
{
// digits returns non-sign bits, so add 1 and divide by 8 (bits in a byte)
return (std::numeric_limits<int>::digits+1)/8;
}
Pointer subtraction in C++ gives the number of elements between
the pointed to objects. In other words, intPtr2 - intPtr1
would return the number of int between these two pointers.
The program wants to know the number of bytes (char), so it
converts the int* to char*. Apparently, the author doesn't
want to use reinterpret_cast either. And static_cast will
not allow a direct convertion from int* to char*, so he
goes through void* (which is allowed).
Having said all that: judging from the name of the function and
how the pointers are actually initialized, a much simpler
implementation of this would be:
int
intSize()
{
return sizeof( int );
}
There is actually no need to convert to void*, other than avoiding reinterpret_cast.
Converting from a pointer-to-int to a pointer-to-char can be done in one step with a reinterpret_cast, or a C-style cast (which, by the standard, ends up doing a reinterpret_cast). You could do a C-style cast directly, but as that (by the standard) is a reinterpret_cast in that context, you'd violate the requirements. Very tricky!
However, you can convert from an int* to a char* through the void* intermediary using only static_cast. This is a small hole in the C++ type system -- you are doing a two-step reinterpret_cast without ever calling it -- because void* conversion is given special permission to be done via static_cast.
So all of the void* stuff is just to avoid the reinterpret_cast requirement, and would be silly to do in real code -- being aware you can do it might help understanding when someone did it accidentally in code (ie, your int* appears to be pointing at a string: how did that happen? Well, someone must have gone through a hole in the type system. Either a C-style cast (and hence a reinterpret_cast), or it must have round-tripped through void* via static_cast).
If we ignore that gymnastics, we now have an array of int. We take pointers to adjacent elements. In C++, arrays are packed, with the difference between adjacent elements equal to the sizeof the elements.
We then convert those pointers to pointers-to-char, because we know (by the standard) that sizeof(char)==1. We subtract these char pointers, as that tells us how many multiples-of-sizeof(char) there are between them (if we subtract int pointers, we get how many multiples-of-sizeof(int) there are between them), which ends up being the size of the int.
If we try to print charPtr1 through std::cout, std::cout assumes that our char* is a pointer-to-\0-terminated-buffer-of-char, due to C/C++ convention. The first char pointed to is \0, so std::cout prints nothing. If we wanted to print the pointer value of the char*, we'd have to cast it to something like void* (maybe via static_cast<void*>(p)).
Please read this: richly commented.
int intSize()
{
int intArray[2]; // Allocate two elements. We don't need any more than that.
/*intPtr1 and intPtr2 point to the addresses of the zeroth and first array elements*/
int* intPtr1 = &intArray[0]; // Arrays in C++ are zero based
int* intPtr2 = &intArray[1];
/*Note that intPtr2 - intPtr1 measures the distance in memory
between the array elements in units of int*/
/*What we want to do is measure that distance in units of char;
i.e. in bytes since once char is one byte*/
/*The trick is to cast from int* to char*. In c++ you need to
do this via void* if you are not allowed to use reinterpret_cast*/
void* voidPtr1 = static_cast<void*>(intPtr1);
char* charPtr1 = static_cast<char*>(voidPtr1);
void* voidPtr2 = static_cast<void*>(intPtr2);
char* charPtr2 = static_cast<char*>(voidPtr2);
/*The distance in memory will now be measure in units of char;
that's how pointer arithmetic works*/
/*Since the original array is a contiguous memory block, the
distance will be the size of each element, i.e. sizeof(int) */
return charPtr2 - charPtr1;
}