Re: Invalid comparator for STL map solution - C / C++ / MFC Discussion Boards

Re: Invalid comparator for STL map

Daniel Pfeffer25-Dec-22 19:36

Daniel Pfeffer

25-Dec-22 19:36

Shouldn't you be passing usingrange as a reference to your comparison function?

C++

bool operator<(const usingrange& x) const

Freedom is the freedom to say that two plus two make four. If that is granted, all else follows.
-- 6079 Smith W.

Re: Invalid comparator for STL map

ForNow25-Dec-22 21:33

ForNow

25-Dec-22 21:33

It’s a member of key type

Re: Invalid comparator for STL map

ForNow26-Dec-22 5:14

ForNow

26-Dec-22 5:14

this is were I am getting the exception at the line with STL_VERIFY

C++

// FUNCTION TEMPLATE _Debug_lt_pred
template <class _Pr, class _Ty1, class _Ty2,
    enable_if_t<is_same_v<_Remove_cvref_t<_Ty1>, _Remove_cvref_t<_Ty2>>, int> = 0>
constexpr bool _Debug_lt_pred(_Pr&& _Pred, _Ty1&& _Left, _Ty2&& _Right) noexcept(
    noexcept(_Pred(_Left, _Right)) && noexcept(_Pred(_Right, _Left))) {
    // test if _Pred(_Left, _Right) and _Pred is strict weak ordering, when the arguments are the cv-same-type
    const auto _Result = static_cast<bool>(_Pred(_Left, _Right));
    if (_Result) {
        _STL_VERIFY(!_Pred(_Right, _Left), "invalid comparator");

this is my code

C++

struct usingrange
{
	ESDID esdid;
	int start;
	int end;
	
	usingrange() { start = 0; end = 0; return; }
  bool operator<(const usingrange& x) const {
	  if (esdid < x.esdid)
		  return esdid < x.esdid;
	  if (start < x.start) 
		  return start < x.start;
	  if (end < x.end) return end < x.end;
	  else
		  return x.end < end;

when I change mt code to

C++

bool operator<(const usingrange& x) const {

    return start < x.start;
 }

everything works I dont get it the operator is excepting a bool retrun code true or false in either I retrun that

Re: Invalid comparator for STL map

Graham Breach26-Dec-22 6:15

Graham Breach

26-Dec-22 6:15

Does your ESDID class have an operator<() function? If not, add one.

Re: Invalid comparator for STL map

ForNow26-Dec-22 6:51

ForNow

26-Dec-22 6:51

Thank you so much actually everything is a structure I am a very open Smile | :)

person

I got to tell you something you C++ guys a geniuses

Re: Invalid comparator for STL map

ForNow26-Dec-22 7:39

ForNow

26-Dec-22 7:39

still getting the exception if the key to the map is a structure

consisting of a char[4] and 2 int how would I set < operator

the only way I got it to work

was by return int1 < x.int1

what would like to do

is concatenate all of them in 12 bytes and compare 12 bytes against 12 bytes
thanks

Re: Invalid comparator for STL map solution

ForNow26-Dec-22 8:31

ForNow

26-Dec-22 8:31

Since my data was 12 bytes long I was lucky there happened to be a data type 16 bytes long a "long double" I just filled the first 4 bytes with zeros and

C++

;
struct usingrange
{
	ESDID esdid;
	int start;
	int end;
	
	usingrange() { start = 0; end = 0; return; }
  bool operator<(const usingrange& x) const {
	  union
	  {
		  long double a;
		  char str1[16];
	  }c;
	  union
	  {
		  long double b;
		  char str2[16];
	  }d;
	
	  ::memset(&c.str1, 0x00, 16);
	  ::memset(&d.str2, 0x00, 16);
	  ::memcpy(&c.str1[4], &esdid, 4);
	  ::memcpy(&c.str1[8],&start,4);
	  ::memcpy(&c.str1[12], &end, 4);
	  ::memcpy(&d.str2[4], &x.esdid, 4);
	  ::memcpy(&d.str2[8], &x.start, 4);
	  ::memcpy(&d.str2[12], &x.end, 4);

	  return c.a < d.b;

	//	  return usingrange  < x;
	//	  if (start < x.start) 
     //   return start < x.start;
//		  return x.end < end;

 }
 
};

Re: Invalid comparator for STL map solution

k505426-Dec-22 15:49

k5054

26-Dec-22 15:49

That has a couple of issues. Firstly, you are invoking undefined behavior: Union declaration - cppreference.com
Secondly, you're expecting that the bit patterns of 2 doubles can be compared. This may not be the case, as you may a NaN this way, and then any comparison will return false:

Shell

[k5054@localhost tmp]$ cat example.c 
#include <stdio.h>

int main()
{
    double d = 0.0 / 0.0;  // produces NaN 
    printf("1.0 < d = %d\n", 1.0 < d);
    printf("d < 1.0 = %d\n", d < 1.0);
    printf("d == d  = %d\n", d == d);
}
[k5054@localhost tmp]$ gcc example.c -o example
[k5054@localhost tmp]$ ./example 
1.0 < d = 0
d < 1.0 = 0
d == d  = 0
[k5054@localhost tmp]$

As we can see in the above code, NaN compared to anything, even itself returns false.

If all you want to do is compare bits, you might as well just go ahead and use memcmp()

C++

*#35;include <cstring>
struct usingrange {
    ESDID esdid;
    int start;
    int end;
    // ... other stuff ...
    bool operator<(const usingrange& x)const {
        return ::memcmp(this, &x, sizeof(*this)
    }
};

This relies on the fact that the sizeof a struct or a class is the size of its individual members. This only becomes tricky if your struct/class includes STL or other objects like std::string or std::vector.

Keep Calm and Carry On

Re: Invalid comparator for STL map solution

ForNow26-Dec-22 21:06

ForNow

26-Dec-22 21:06

The return using memcmp as operands will get an exception in order to get around this I need to to have to have the syntax return operand1 < operand2 truthfully the order of the items is not that important as long as I am able to retrieve the data by key thanks

Re: Invalid comparator for STL map solution

Graham Breach26-Dec-22 21:42

Graham Breach

26-Dec-22 21:42

I think you're making it too complicated:

C++

struct ESDID
{
  char c[4];

  bool operator<(const ESDID& other) const
  {
    return c[0] < other.c[0] && c[1] < other.c[1] &&
      c[2] < other.c[2] && c[3] < other.c[3];
  }
}

struct usingrange
{
  ESDID esdid;
  int start;
  int end;

  bool operator<(const usingrange& other) const
  {
    return esdid < other.esdid && start < other.start &&
      end < other.end;
  }
}

I've used && and < because you do not want operator<() to return true for two different objects when you swap them around.

Re: Invalid comparator for STL map solution

ForNow27-Dec-22 1:22

ForNow

27-Dec-22 1:22

just tried it that did it thank you Smile | :)

Read Access exception in xtree for map::insert

ForNow24-Dec-22 17:18

ForNow

24-Dec-22 17:18

hi

I am getting a read access error in xtree line 1607. This after i have inserted 4,527 map entries successfully I did a call to map::maxsize using a unsigned integer and it came up with a number 0x76767676 I dont think memory is the issue as I have 64 gb machine

I bypassed the entry just to see if there was something wrong with that entry and go an error on the next as well

here is the microsoft code I underlined the while statement with the issue intellesense is giving the message _Trnode was 0x40D51F000040.

any guidence who i shoud go about debugging would be appreciated

C++

template <class _Keyty>
    _Tree_find_result<_Nodeptr> _Find_lower_bound(const _Keyty& _Keyval) const {
        const auto _Scary = _Get_scary();
        _Tree_find_result<_Nodeptr> _Result{{_Scary->_Myhead->_Parent, _Tree_child::_Right}, _Scary->_Myhead};
        _Nodeptr _Trynode = _Result._Location._Parent;
        while (!_Trynode->_Isnil) {
            _Result._Location._Parent = _Trynode;
            if (_DEBUG_LT_PRED(_Getcomp(), _Traits::_Kfn(_Trynode->_Myval), _Keyval)) {
                _Result._Location._Child = _Tree_child::_Right;
                _Trynode                 = _Trynode->_Right;
            } else {
                _Result._Location._Child = _Tree_child::_Left;
                _Result._Bound           = _Trynode;
                _Trynode                 = _Trynode->_Left;
            }
        }

Re: Read Access exception in xtree for map::insert answer found

ForNow25-Dec-22 9:55

ForNow

25-Dec-22 9:55

I had a buffer that was way too small it overlayed data inluding the pointer to the map class

thanks

overloaded = operator not being invoked

ForNow22-Dec-22 14:36

ForNow

22-Dec-22 14:36

Hi

my overloaded operator are in my header so I make a breakpoint watching them being invoked except for the = operator hope some can tell me why

i'll post

first type of struct using range

C++

;
struct usingrange
{
	ESDID esdid;
	int start;
	int end;
	uint32_t operator=(const BYTE z[4])
	{
		return ((z[0] << 24) | (z[1] << 16) | (z[2] << 8) | z[3]);
	
	};

  bool operator<(const usingrange x) const {
	  if (esdid < x.esdid)  return esdid < x.esdid;
	  if (start < x.start) return start < x.start; if (end < x.end) return end < x.end;
  };
 
};

next my code with the = operator the lvalue is member of using range I would think that would casue invocation of my overloaded operator

C++

usingrange using_range{ usingpoint->usingesdid };

using_range.start = (BYTE &)usingpoint->statement;
using_range.end = (BYTE &)usingpoint->laststatement;

Re: overloaded = operator not being invoked

Mircea Neacsu22-Dec-22 15:24

Mircea Neacsu

22-Dec-22 15:24

Again, not sure which one of the 3 statements you would expect to invoke the assignment operator. The first one:

C++

usingrange using_range{ usingpoint->usingesdid };

is a declaration that invokes a constructor. As you don't have a constructor, the compiler creates a default one.

The second and the third:

C++

using_range.start = (BYTE &)usingpoint->statement;
using_range.end = (BYTE &)usingpoint->laststatement;

are simply integer assignments.

If you would want to invoke the assignment operator you would need to write something like

C++

usingrange using_range2;
using_range2 = using_range;

Mircea

Re: overloaded = operator not being invoked

ForNow22-Dec-22 15:28

ForNow

22-Dec-22 15:28

these two the lvalue is a int and a member of using range the rvalue is of type BYTE [4] ? I am sure I am not understanding something ?

C++

using_range.start = *usingpoint->statement;
	    using_range.end = *usingpoint->laststatement;

Re: overloaded = operator not being invoked

Mircea Neacsu22-Dec-22 15:37

Mircea Neacsu

22-Dec-22 15:37

If statement is BYTE[4], usingpoint->statement is a pointer to a BYTE value. So *usingpoint->laststatement is the content of the first byte. It can be assigned to an integer without any problem. That's probably not what you want but that's what the compiler understood.

The assignment operator that you defined is for assigning the WHOLE structure. It doesn't apply to individual members.

Mircea

Re: overloaded = operator not being invoked

ForNow22-Dec-22 16:00

ForNow

22-Dec-22 16:00

you are right I saw that in that disassembly however I made a breakpoint at "int operator+..." and it never got invoked probably because as you said the rvalue is BYTE while in the operator= parameter it BYTE z[4]; when I have the code

C++

using_range.start = usingpoint->statement;

I got an error message from intellsense saying cannt assign int to BYTE *

Re: overloaded = operator not being invoked

Mircea Neacsu22-Dec-22 16:32

Mircea Neacsu

22-Dec-22 16:32

I think you misunderstood the functioning of overloaded operators. I'll try to explain but it's best if you check with a proper manual. I'm not a particularly good teacher Smile | :)

If you have something like:

C++

struct A {
  char ch;
  long l;
//...
};

struct B {
  int i;
  char c;
//...
};

A a;
B b;

If you write an assignment:

C++

a = b;

The compiler will complain that it doesn't know how to assign a B to an A. Kind of an apples and oranges problem.
However you can write:

C++

a.l = b.i;
a.ch = b.c;

because those are standard types and compiler knows how to perform those assignments without needing a user defined assignment operator.

Now we add a user defined assignment operator:

C++

struct A {
  char ch;
  long l;
//...
  int operator =(B& rhs) {l = rhs.i; ch = rhs.c; return 1;}
};

After that we can write:

C++

a = b;

and the compiler will know how to do the assignment.

You can have many such assignment operators. For instance you might add another one to assign an A to another A:

C++

struct A {
  char ch;
  long l;
//...
  int operator =(B& rhs) {l = rhs.i; ch = rhs.c; return 1;}
  int operator =(A& rhs) {l = rhs.l; ch = rhs.ch; return 1}
};

The assignment operator that takes an object of the same kind on the right hand side is called "principal assignment operator".

Note however that we cannot "chain" the assignment operator. If we write:

C++

A a1;
a1 = a = b;

we will get an error because the result of

C++

a = b

is an int and cannot be assigned to an A. To respect the conventions we would have to change the signature of the principal assignment operator:

C++

struct A {
  char ch;
  long l;
//...
  int operator =(const B& rhs) {l = rhs.i; ch = rhs.c; return 1;}
  A& operator =(const A& rhs) {l = rhs.l; ch = rhs.ch; return *this;}
};

Note that I also made the right hand side a reference to a const.

I hope you will find useful my little lecture on assignment operators but, again, best would be to check with a real manual.

Mircea

Re: overloaded = operator not being invoked

ForNow22-Dec-22 16:41

ForNow

22-Dec-22 16:41

thanks beginning to see there is not a way to do what I want with operator overloading

Still Struggling with map::insert and operator overload

ForNow20-Dec-22 15:57

ForNow

20-Dec-22 15:57

Hi
I am still having problems with map::insert

Here is my data

C++

struct usingx
{
	unsigned char rectype;
#define using 0
#define pop 0x20
#define push 0x40
#define drop 0x80
	unsigned char usingflag
#define oridinaryusing 0
#define labeledusing 0x10
#define depedentusing 0x20
#define labeldependentusing 0x30
		;	ESDID locesdid;
	BYTE statement[4];
	BYTE location[4];
	BYTE usingvalue[4];
	BYTE laststatement[4];
	ESDID usingesdid;
	unsigned char registerx;
	BYTE displacement[2];
	BYTE reservedz;
	BYTE usingrange[4];
	BYTE reservedu[2];
	BYTE labeloffset[4];
	BYTE labelength[4];
	char* labelx;


};

Here is the key that I am using the start and end variable are big endian so I have to convert it and overloaded the = operator;

C++

struct usingrange
{
	ESDID esdid;
	int start;
	int end;
	int operator=(BYTE z[4]) { return _byteswap_ulong((int)z[4]); };

  bool operator<(const usingrange x) const {
	  if (esdid < x.esdid)  return esdid < x.esdid;
	  if (start < x.start) return start < x.start; if (end < x.end) return end < x.end;
  };

here is the code using the map::insert

C++

usingrange using_range{ usingpoint->usingesdid };
using_range.start = *usingpoint->statement;
using_range.end = *usingpoint->laststatement;

procpointer->usingptr->insert({ using_range,*usingpoint });
break;

I made breakpoints on the operator= code and thought when I stepped into it with the vs debugger it would trace it but it didnt

both start and end ended up zeros

dont know why ?

Re: Still Struggling with map::insert and operator overload

Mircea Neacsu20-Dec-22 17:34

Mircea Neacsu

20-Dec-22 17:34

The line

C++

usingrange using_range{ usingpoint->usingesdid };

does not invoke your assignment operator. This is a constructor and it will just copy the argument. Not sure where you expected the assignment to be invoked.

I would replace the assignments with a constructor similar to this:

C++

 struct usingrange
{
    usingrange(usingx& x) {
      esdid = byte_swap(x.usingesdid;
      start = x.statement;
      end = x.laststatement;
      //…
    }
	ESDID esdid;
	int start;
	int end;

As a comment, your comparison operator is suboptimal and incomplete. I’m surprised you didn’t get a warning that not all control paths return a value.

Mircea

Re: Still Struggling with map::insert and operator overload

ForNow21-Dec-22 1:51

ForNow

21-Dec-22 1:51

I understand at this point I am trying to get the =operator overload working I think it’s only referencing the first byte as opposed to 4 bytes
Thought if I cast it it would work but didn’t going to try to move it to a int

Message Closed

18-Dec-22 5:01

ForNow

18-Dec-22 5:01

Message Closed

modified 18-Dec-22 13:49pm.

Re: Question about return value of STD MAP::INSERT

Mircea Neacsu18-Dec-22 6:55

Mircea Neacsu

18-Dec-22 6:55

You can simplify to:

C++

auto ret =  procpointer->extsymcollector->insert({exsympointer->SYMESDID, *exsympointer})

ForNow wrote:
my code does NOT complie

What compiler to you use? what message to you get?

Mircea

Use Ctrl+Left/Right to switch messages, Ctrl+Up/Down to switch threads, Ctrl+Shift+Left/Right to switch pages.