Error when setting 2D array to 2D array

Question

I have this constructor:

Transform::Transform( float matrix[4][4] )
{
    m = matrix;
}

And this class definition:

class Transform
{
    float m[4][4];
public:
    Transform();
    Transform(float matrix[4][4]);

But this does not compile.

What could be wrong?

Error 1 error C2440: '=' : cannot convert from 'float [][4]' to 'float [4][4]' c:\Users\Josh\Documents\agui\trunk\src\Agui\Transform.cpp 75

Thanks

I'd recommend std::array if you can use it. It does assignment. — Qaz
– Qaz, Commented Jan 9, 2013 at 2:06
Arrays are not copyable in this manner. You need to copy the contents (either through indexed access or a memory-move algorithm like std::copy(), memcpy() or memmove(). Of course, belay all that and do what chris says above. Use your stdlib. Its whats for dinner. — WhozCraig
– WhozCraig, Commented Jan 9, 2013 at 2:07

Karthik T · Accepted Answer · 2013-01-09 02:20:49Z

4

If you are using c++11 try to change float matrix[4][4] to std::array<std::array<float,4>,4>

It is a mouthful, but it supports such operations that c arrays do not natively support.

You could do something like this to clean up the syntax.

typedef std::array<std::array<float,4>,4> Matrix;

Now you can do

Matrix myMatrix;

p.s If you are not using C++11, you could use vector instead of array. It is a little different from array, but adds more features as well, and after you set it up access is identical.

edited Jan 9, 2013 at 2:20

answered Jan 9, 2013 at 2:08

Karthik T

32.1k7 gold badges71 silver badges92 bronze badges

Sign up to request clarification or add additional context in comments.

3 Comments

Pubby Over a year ago

With a typedef so it isn't a mouthful.

Qaz Over a year ago

@Pubby, It's C++11 anyway. template<typename T> using array4 = std::array<T, 4>; :) Depends on what changes when you use it, really.

Karthik T Over a year ago

@Pubby very true, I have added it

David Venegoni · Accepted Answer · 2013-01-09 02:11:09Z

3

Karthik's answer is excellent, alternatively, you could also do...

  for(int i = 0; i < 4; i++)
  {
     for(int j = 0; j < 4; j++)
      {
         m[i][j] = matrix[i][j];
      }
  }

The principle is the same that WhozCraig mentioned in the comment.

answered Jan 9, 2013 at 2:11

David Venegoni

5083 silver badges14 bronze badges

Comments

Code-Apprentice · Accepted Answer · 2013-01-09 19:00:32Z

1

Even though you declare the parameter to your constructor as float matrix[4][4], the compiler ignores the first 4.

edited Jan 9, 2013 at 19:00

answered Jan 9, 2013 at 2:14

Code-Apprentice

84k26 gold badges163 silver badges290 bronze badges

4 Comments

jmasterx Over a year ago

I guess then it sees it as float*[4]

Code-Apprentice Over a year ago

@Milo You will need to ask a language Nazi for the difference between float [][4] (as in your error message) and float *[4]. I typically treat them as the same thing (which probably isn't technically accurate).

WhozCraig Over a year ago

Not a lang-nazi, but so you know , float(*)[4]. With a variable decl it would be float (*ar)[4]. And its the first dimension (the superior dim) that is ignored; not all but the last. The sub-dimensions are used to calculate linear offset into the backdrop of the memory behind the array.

Code-Apprentice Over a year ago

@WhozCraig Thanks for the clarification. I removed that sentence from my answer.

Collectives™ on Stack Overflow

Error when setting 2D array to 2D array

3 Answers 3

3 Comments

Comments

4 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

3 Comments

Comments

4 Comments

Your Answer

Sign up or log in

Post as a guest

Related