C help needed

[aneeshm]

For some reason, this code simply isn't working.

#include
#include
#include

/* float determinant(float *a, int n) {{{1 */
float determinant(float *a, int n)
{
	float *b, det=0;

	if(n<2)
		return 0;

	if(n==2)
		det=( (*a)*(*(a+3)) - (*(a+1))*(*(a+2)));
	else
	{
		b=malloc((n-1)*(n-1)*sizeof(float));

		int i,j,k,m=0,l=0;

		for(i=0;i

The errors it is giving:

*** glibc detected *** ./a.out: free(): invalid next size (fast): 0x0000000000601040 ***
======= Backtrace: =========
/lib/libc.so.6[0x2ab41ff97b23]
/lib/libc.so.6(cfree+0x8c)[0x2ab41ff9b26c]
./a.out[0x400877]
./a.out[0x40095a]
/lib/libc.so.6(__libc_start_main+0xf4)[0x2ab41ff458e4]
./a.out[0x4005e9]
======= Memory map: ========
00400000-00401000 r-xp 00000000 08:04 4014120                            /home/aneeshm/work/se/pl-ass/mat/a.out
00600000-00601000 rw-p 00000000 08:04 4014120                            /home/aneeshm/work/se/pl-ass/mat/a.out
00601000-00622000 rw-p 00601000 00:00 0                                  [heap]
2ab41fa89000-2ab41faa5000 r-xp 00000000 08:04 8880137                    /lib/ld-2.5.so
2ab41faa5000-2ab41faaa000 rw-p 2ab41faa5000 00:00 0 
2ab41fca4000-2ab41fca6000 rw-p 0001b000 08:04 8880137                    /lib/ld-2.5.so
2ab41fca6000-2ab41fd27000 r-xp 00000000 08:04 8912908                    /lib/libm-2.5.so
2ab41fd27000-2ab41ff26000 ---p 00081000 08:04 8912908                    /lib/libm-2.5.so
2ab41ff26000-2ab41ff28000 rw-p 00080000 08:04 8912908                    /lib/libm-2.5.so
2ab41ff28000-2ab42006f000 r-xp 00000000 08:04 8880190                    /lib/libc-2.5.so
2ab42006f000-2ab42026f000 ---p 00147000 08:04 8880190                    /lib/libc-2.5.so
2ab42026f000-2ab420272000 r--p 00147000 08:04 8880190                    /lib/libc-2.5.so
2ab420272000-2ab420274000 rw-p 0014a000 08:04 8880190                    /lib/libc-2.5.so
2ab420274000-2ab42027a000 rw-p 2ab420274000 00:00 0 
2ab42027a000-2ab420287000 r-xp 00000000 08:04 8880139                    /lib/libgcc_s.so.1
2ab420287000-2ab420487000 ---p 0000d000 08:04 8880139                    /lib/libgcc_s.so.1
2ab420487000-2ab420488000 rw-p 0000d000 08:04 8880139                    /lib/libgcc_s.so.1
2ab424000000-2ab424021000 rw-p 2ab424000000 00:00 0 
2ab424021000-2ab428000000 ---p 2ab424021000 00:00 0 
7fff8b00c000-7fff8b021000 rw-p 7fff8b00c000 00:00 0                      [stack]
ffffffffff600000-ffffffffff601000 r-xp 00000000 00:00 0                  [vdso]
Aborted (core dumped)

I've not been able to make head or tail of it.

EDIT: For some reason, the "code" markers don't play nice with C. So I'll attach the file in the next post.

[Kuciwalker]

Try finishing that last line of code and closing your braces

And fix the includes?

edit: oh, the carets must be screwing stuff up.

[aneeshm]

Originally posted by Kuciwalker
Try finishing that last line of code and closing your braces

And fix the includes?

[aneeshm]

Can't even upload a damn C file.

[Kuciwalker]

Do you know where in the code the error occurs? (Which free?)

[Kuciwalker]

Oh, and after a couple minutes I have no idea what's wrong.

[aneeshm]

No. I'll try to find out.

[aneeshm]

Originally posted by Kuciwalker
Oh, and after a couple minutes I have no idea what's wrong.

As far as I can make out, the thing should work fine.

[aneeshm]

A quick GDB session tells me that it's breaking the first time free is called within the determinant function.

[VetLegion]

You crazy Indian, you are writing to memory that isn't yours!

This offset:

(n-1)*l + m

Is going to be larger than the

(n-1)*(n-1)

you are allocating.

You can solve this much more elegantly without pointers BTW.

[aneeshm]

OK, thanks. I got it. I forgot to initialise l and m to 0 for each iteration of the i loop.

*Sigh*

It's always, ALWAYS the silly, trivial, moronic mistakes that make the for worst debugging nightmares.

[VetLegion]

OK, I happen to have a Borland compiler around here, I'll compile this for you to see if I was right...

[aneeshm]

And what's the method without pointers?

[aneeshm]

Originally posted by VetLegion
OK, I happen to have a Borland compiler around here, I'll compile this for you to see if I was right...

I just checked.

It works.

And it compiles fine with gcc, you don't really need a Borland compiler.

[VetLegion]

Well the natural way would be array, saving you some calculation and being easier to read and slightly harder to mess up. I have to run, maybe I'll come back to this later and we can discuss your code