Passing 2D array with variable dimensions as function argument

Question

I just saw the following code among the successful submissions at codechef.

http://www.codechef.com/viewplaintext/1595846

I used to think that

float max(int n,int arr[n][n])
{....}

is not allowed in C++ (as 'n' is a variable). My CodeBlocks (on windows) with MinGW [gcc 4.4] gives compile time error. that "error: array bound is not an integer constant.

Then how can be such a solution be accepted by CodeChef's judge. Is there any special flag that allows us to do that in C++???

EDIT: A link showing status as AC (accepted) : http://www.codechef.com/viewsolution/1595846

See Abhishek Thakur answer below. This submission is marked as C, not C++, and this happens to be legal in C (though not in standard C++). — Gorpik
– Gorpik, Commented Dec 11, 2012 at 11:24

Abhishek Thakur · Accepted Answer · 2012-12-11 11:12:29Z

5

Variable-length automatic arrays are allowed in ISO C99, and as an extension GCC accepts them in C90 mode and in C++. These arrays are declared like any other automatic arrays, but with a length that is not a constant expression.

Ref: http://gcc.gnu.org/onlinedocs/gcc/Variable-Length.html

answered Dec 11, 2012 at 11:12

Abhishek Thakur

17.1k16 gold badges71 silver badges98 bronze badges

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

4 Comments

Gorpik Over a year ago

Indeed, the submission says it is implemented in C, not C++. So the code is correct according to the C standard.

TheCrazyProgrammer Over a year ago

Thanks. I tried the same code in C, that worked. But what about C++ ? It is giving errors.

Gorpik Over a year ago

@TheCrazyProgrammer You have errors because this is legal C, but not legal C++. C++98 was almost a strict superset of C90, but C99 came later and introduced some additional features such as this. Later versions of the C++ standard have not made a priority of keeping compatibility with C.

John Bode Over a year ago

Note that, with the C2011 standard, variable-length arrays are now optional; it's up to the implementation do decide whether to support them or not. GCC will probably continue to do so, but compilers that currently don't may choose not to going forward. And yes, C and C++ are different languages, with different semantics, and will probably diverge even further over time. Do not assume any arbitrary C code will compile as C++.

Archy · Accepted Answer · 2012-12-11 11:34:57Z

4

I stand corrected: C99 does allow this for C, although many compilers do not implement it yet and some probably never will (microsoft).

Previous answer

Either pass arr as int** or use something like

template< int N >
float max(const int (&arr)[N][N])
{ ... }

which off course requires N to be a compile time constant. The safest solution would be to use a std::vector or some other container that has knowledge about its size.

Overall the code seems pretty fragile to me.

edited Dec 11, 2012 at 11:34

answered Dec 11, 2012 at 10:49

Archy

2061 silver badge6 bronze badges

Collectives™ on Stack Overflow

Passing 2D array with variable dimensions as function argument

2 Answers 2

4 Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

4 Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related